TSMC’s Arizona Plant Matches Taiwan’s Chip Yields, Moving from Open Source to Proprietary License Is Not Always a Means to Financial Growth, Salesforce Shakes Up AI Pricing with $2 Per Conversation Model
All this and more on the “The price is right” edition of Hashtag Trending. I’m your host, Jim Love. Let’s get into it.
In a significant development for the US semiconductor industry, Taiwan Semiconductor Manufacturing Co. (TSMC) has reportedly achieved production yields at its new Arizona facility that match those of its established plants in Taiwan. This milestone comes from recent trial production runs using advanced 4-nanometer process technology, according to an anonymous source familiar with the company’s operations.
The $65 billion Arizona project, consisting of three fabs, represents the largest foreign direct investment in both Arizona’s history and in a US greenfield project. It’s a crucial part of the US government’s efforts to bolster domestic chip manufacturing, with TSMC set to receive up to $6.6 billion in grants and $5 billion in loans for the initiative.
Despite initial delays due to workforce challenges and cultural differences, TSMC’s success in matching Taiwanese yield rates is a positive sign for the project’s viability. The first fab is now scheduled to begin mass production in the first half of 2025, with the second and third fabs focusing on even more advanced 2nm and 3nm processes in the following years.
This development is particularly noteworthy as it addresses previous concerns about whether TSMC could maintain its high efficiency standards outside of Taiwan. With comparable yields, the company should be able to maintain its targeted gross margin rates of 53 percent or higher in its US operations.
As the global chip shortage continues to impact various industries, TSMC’s progress in Arizona represents a significant step towards enhancing the resilience of the US semiconductor supply chain. It also demonstrates the potential for successful high-tech manufacturing collaborations between the US and its Asian allies.
Sources include: TechSpot (https://www.techspot.com/news/100849-tsmc-65-billion-arizona-facility-can-now-match.html), Bloomberg
In light of Elastics move to a proprietary license, there is a small but real trend of companies switching their software from open source to proprietary licenses.
But does this strategy actually improve their bottom line? It’s not that clear cut.
Rachel Stephens of RedMonk examined the financial outcomes of four major tech companies – MongoDB, Elastic, HashiCorp, and Confluent – all of which transitioned from open source to proprietary licensing models. The study focused on changes in revenue, market capitalization, and net income before and after the license changes.
The results paint what can only be called a nuanced picture. While all four companies saw revenue growth after changing licenses, the rate of increase wasn’t significantly different from their pre-change trajectory. This suggests that license changes alone may not be the revenue booster some companies expect.
Market capitalization told a more varied story. MongoDB experienced substantial growth post-license change, Elastic saw moderate gains, but HashiCorp actually lost value. These mixed results indicate that changing to a proprietary license doesn’t guarantee an increase in company valuation.
Interestingly, none of the companies studied have yet achieved profitability, with their valuations primarily driven by expected future growth. This underscores the complex factors at play in tech company valuations beyond just licensing models.
But there are other issues that might affect the end users of the software and the open source community in general.
There’s always been a suspicion about converting an open-source company to a proprietary model. Decades ago, companies like SugarCRM who had leveraged a community of developers for their initial development and market penetration, moved to a proprietary model and some felt, abandoned the people who had built the software and essentially allowing corporate and other “investors” to be able to profit from their innovation and labour.
Recently, IBMs acquisition of Red Hat ultimately led to friction with the rest of the open-source Linux community.
But the recent changes, may stem from another rationale. Elastic’s recent shift to proprietary licensing was that public cloud providers could offer their open-source version “as a service.” So these open-source companies were providing free software to giant tech companies who may or may not have contributed anything to its development.
Stephens’ report really doesn’t look at them impact on the open-source world, but it does take some first steps in perhaps initiating a larger discussion about the future of open-source.
This conversion to proprietary software is only one aspect, but it is, according to this report, a strategy that’s gaining traction, especially in the database sector.
And the report indicates that in the debate over open source versus proprietary licensing continues, this study suggests that companies should carefully consider their motivations for license changes. While it may seem like a straightforward path to improved financials, the reality appears far more complex.
Sources include: RedMonk (https://redmonk.com/rstephens/2024/08/26/software-licensing-changes-and-their-impact-on-financial-outcomes/)
In a move that could reshape the AI industry’s pricing landscape, Salesforce has announced a cost structure for its upcoming AI agent. CEO Marc Benioff revealed that the company plans to charge just $2 per conversation for their AI tool, potentially setting a new standard in the rapidly evolving market for enterprise AI solutions.
This pricing strategy comes as Benioff takes aim at competitor Microsoft, claiming that their generative AI product, Copilot, has left many customers underwhelmed. By offering a pay-per-use model at such a competitive rate, Salesforce is positioning itself to capture a significant share of the growing demand for AI-powered business tools.
The $2 per conversation pricing could have far-reaching implications for the industry. It may pressure other major players to reconsider their pricing models, potentially leading to more accessible AI tools across the board.
For businesses, this could mean more flexibility in integrating AI into their operations without committing to fixed subscription fees.
However, questions remain about this model. At $2 per call, it’s likely to be competitive or a cost saving if compared with a skilled agent, provided it can provide an equal level of service and replace at least some operators. That would be a sophisticated system. If it turns out to be the equivalent what could be handled by a simple AI agent, it could end up being a more costly alternative.
There’s a real danger, as we have noted before in the “death of a thousand paper cuts” with SaaS software where each little application is a small monthly charge. For many SaaS options it’s a management nightmare of keeping track of all the subscriptions – are they being used? Or do we have virtual shelfware? Did we cancel when people left or just stopped using the SaaS products.
Salesforce made its fortune pushing SaaS software that had a low monthly entry price, but made its big money by selling you add-ons. At one point, long ago, we estimated that the per user price that looked so attractive in the initial conversations would double or even close to triple by the time we got all the add-ons we actually needed.
There are have been horror stories of SaaS implementations that ended up costing a lot more than anyone thought possible. And of course, there are the potential cost increases – witness what happened recently with Canva where some users reported as being as high as 300%. This was supposedly the bargain replacement for the more expensive Adobe suite.
And those who think that all you have to do is to not renew if prices are too high, may realize that there is more to shifting software products than simply signing with a new partner. Data conversion, training and changing your processes could be a huge barrier to change.
Finally, while new pricing models are novel, there is a reason why so many people have been upset with cell phone companies – the different pricing models made comparison difficult and sometime impossible.
Companies are investing millions and even potentially billions in adding AI to their offerings. They’ll want that back, with a profit.
So – this is a classic case of caveat emptor. Do a careful business case.
Be prepared to build pricing models. And make absolutely certain that you know what the implications are on multiple scenarios.
As the AI arms race heats up among tech giants, Salesforce’s bold pricing move could be a game-changer, or it could be an embarrassment as you explain to your CFO where the savings went.
Sources include: Analytics India Magazine (https://analyticsindiamag.com/ai-breakthroughs/salesforce-pricing-strategy-could-dictate-industry-trend-for-ai-agents/)
And that’s our show for today. You can find show notes at our news site technewsday.com or .ca take you pick.
Thanks for listening. I’m your host Jim Love, have a Terrific Tuesday.
