Offshore
Video
Michael Fritzell (Asian Century Stocks)
Recent write-ups on Asian Century Stocks
Click to see the full list: https://t.co/FmCgd1HmEy https://t.co/nofJvCdsyD
tweet
Recent write-ups on Asian Century Stocks
Click to see the full list: https://t.co/FmCgd1HmEy https://t.co/nofJvCdsyD
tweet
Michael Fritzell (Asian Century Stocks)
So, an x% efficiency gain in coding, but potential for tech debt for those aren't careful?
tweet
So, an x% efficiency gain in coding, but potential for tech debt for those aren't careful?
Talking to CTOs, there is a worry that junior engineers are producing so much tech debt with AI, they'd rather them code by hand and learn. Junior engs don't know what "good" looks like. Personally, I'm reviewing every line and throw away ~35% of AI code. tweet
X (formerly Twitter)
ARENA MAN CAPITAL (@ArenaManCapital) on X
Talking to CTOs, there is a worry that junior engineers are producing so much tech debt with AI, they'd rather them code by hand and learn. Junior engs don't know what "good" looks like. Personally, I'm reviewing every line and throw away ~35% of AI code.
Offshore
Photo
Michael Fritzell (Asian Century Stocks)
The international market is still small for Nong Shim, but they're building a new export-dedicated plant in Busan expected to complete in the second half of 2026 https://t.co/xWRYsiFDHJ
tweet
The international market is still small for Nong Shim, but they're building a new export-dedicated plant in Busan expected to complete in the second half of 2026 https://t.co/xWRYsiFDHJ
tweet
Offshore
Photo
Michael Fritzell (Asian Century Stocks)
"If your core value proposition can be replicated by an LLM with 90% of the quality at 1% of the cost, and you provide a probabilistic product, you don’t have a sound business model anymore"
https://t.co/y3PBes3zfC
tweet
"If your core value proposition can be replicated by an LLM with 90% of the quality at 1% of the cost, and you provide a probabilistic product, you don’t have a sound business model anymore"
https://t.co/y3PBes3zfC
tweet
Michael Fritzell (Asian Century Stocks)
Probabilistic point solutions (most affected by AI):
- Project management
- Marketing automation
- Workflow tools
- Content generation
- Seat-based models
Deterministic systems of record (least affected):
- Accounting software
- Vertical SaaS eg logistics, insurance systems
tweet
Probabilistic point solutions (most affected by AI):
- Project management
- Marketing automation
- Workflow tools
- Content generation
- Seat-based models
Deterministic systems of record (least affected):
- Accounting software
- Vertical SaaS eg logistics, insurance systems
tweet
Offshore
Photo
Michael Fritzell (Asian Century Stocks)
RT @Aimuso: Which intern vibecoded a release at Substack? https://t.co/BO0QXfPp8c
tweet
RT @Aimuso: Which intern vibecoded a release at Substack? https://t.co/BO0QXfPp8c
tweet
Offshore
Photo
Jukan
Honestly, it's getting embarrassing to even call them the M7 now. Excluding Google, the other M6 stocks are down 5~25% from their all-time highs, and the only thing propping up the M7's gains has been Google.
Over the past 6 months, market cap has increased by $1.7 trillion, right?
tweet
Honestly, it's getting embarrassing to even call them the M7 now. Excluding Google, the other M6 stocks are down 5~25% from their all-time highs, and the only thing propping up the M7's gains has been Google.
Over the past 6 months, market cap has increased by $1.7 trillion, right?
tweet
Michael Fritzell (Asian Century Stocks)
. @ShmuelLon nailed it. At least so far
tweet
. @ShmuelLon nailed it. At least so far
Damn, the rally got killed in Asia. Korea now selling off... tweet
X (formerly Twitter)
Value Drift (@valuedrift) on X
Damn, the rally got killed in Asia. Korea now selling off...
Jukan
BofA (Bank of America) analyst Vivek Arya concluded that it doesn’t make logical sense for the chip sector and the software sector to sell off together. He argued that a decline in chip stocks implies deteriorating AI investment returns (ROI), while a decline in software stocks implies AI adoption will become extremely widespread—so those two conclusions are hard to reconcile at the same time. In his view, the market is overreacting.
Arya noted that the market’s implicit judgment behind the selloff in AI-related chip stocks is that “AI investments have deteriorated to the point where achieving ideal returns is difficult, limiting future growth runway.” By contrast, the selloff in software stocks reflects a different judgment: “AI adoption will be broadly disseminated and meaningfully boost productivity, replacing software businesses and application business models.”
Arya wrote that “these two outcomes cannot occur simultaneously.” He said this positioning resembles the fear seen during the “DeepSeek shock” in January this year—an episode that ultimately proved to lack factual grounding. At the time, the market worried that DeepSeek could build competitive models at lower cost, reducing demand for high-performance chips.
Arya believes what the market actually witnessed after the “DeepSeek shock” was more AI spending and accelerated growth in AI tokens. In his report, he estimated that 2025 cloud capex growth could reach 69%, far above the initial expectation of 20–30%.
He also argued that even if current AI models are highly capable, it may still take years to truly “prove value” from a productivity standpoint—so he does not expect AI investment to slow in the near term. Even if a major productivity breakthrough emerges in the future, models will need continuous improvement to sustain user engagement, meaning investment is unlikely to decline.
His conclusion is that while the ultimate end-state of the software industry is uncertain, the chip industry has already benefited from the AI buildout cycle—and that tailwind should continue. He added that enterprise AI adoption is still early, and more sovereign nations are beginning to increase AI investment. Arya also pointed out that chip-sector valuations already reflect concerns about slowing spend or downward revisions to earnings expectations, but those concerns “may not actually materialize.”
From a trading perspective, this drawdown looks closer to a “narrative-driven correlation shock.” When the market sells software on the thesis that AI will displace it, while simultaneously selling AI chips on the thesis that AI ROI is worsening, core AI assets like Nvidia can easily get dragged down in the process—through position unwinds and broad risk re-rating.
Arya concluded that what Nvidia investors most need in the near term is clarity on whether this price action reflects a real change in AI capex and demand fundamentals, or whether it’s coming from indiscriminate, cross-sector de-risking triggered by the narrative that “AI beats software.”
That was Arya’s interpretation, and I also don’t really understand this selloff.
If I had to find a reason, you could argue that there was simply too much leverage in the system recently, and mechanical selling—like CTA-driven flows—may have contributed.
Still, I think the magnitude of the drop is excessive.
I’m really bummed that I don’t have any cash right now, so I can’t buy more even after this drop.
tweet
BofA (Bank of America) analyst Vivek Arya concluded that it doesn’t make logical sense for the chip sector and the software sector to sell off together. He argued that a decline in chip stocks implies deteriorating AI investment returns (ROI), while a decline in software stocks implies AI adoption will become extremely widespread—so those two conclusions are hard to reconcile at the same time. In his view, the market is overreacting.
Arya noted that the market’s implicit judgment behind the selloff in AI-related chip stocks is that “AI investments have deteriorated to the point where achieving ideal returns is difficult, limiting future growth runway.” By contrast, the selloff in software stocks reflects a different judgment: “AI adoption will be broadly disseminated and meaningfully boost productivity, replacing software businesses and application business models.”
Arya wrote that “these two outcomes cannot occur simultaneously.” He said this positioning resembles the fear seen during the “DeepSeek shock” in January this year—an episode that ultimately proved to lack factual grounding. At the time, the market worried that DeepSeek could build competitive models at lower cost, reducing demand for high-performance chips.
Arya believes what the market actually witnessed after the “DeepSeek shock” was more AI spending and accelerated growth in AI tokens. In his report, he estimated that 2025 cloud capex growth could reach 69%, far above the initial expectation of 20–30%.
He also argued that even if current AI models are highly capable, it may still take years to truly “prove value” from a productivity standpoint—so he does not expect AI investment to slow in the near term. Even if a major productivity breakthrough emerges in the future, models will need continuous improvement to sustain user engagement, meaning investment is unlikely to decline.
His conclusion is that while the ultimate end-state of the software industry is uncertain, the chip industry has already benefited from the AI buildout cycle—and that tailwind should continue. He added that enterprise AI adoption is still early, and more sovereign nations are beginning to increase AI investment. Arya also pointed out that chip-sector valuations already reflect concerns about slowing spend or downward revisions to earnings expectations, but those concerns “may not actually materialize.”
From a trading perspective, this drawdown looks closer to a “narrative-driven correlation shock.” When the market sells software on the thesis that AI will displace it, while simultaneously selling AI chips on the thesis that AI ROI is worsening, core AI assets like Nvidia can easily get dragged down in the process—through position unwinds and broad risk re-rating.
Arya concluded that what Nvidia investors most need in the near term is clarity on whether this price action reflects a real change in AI capex and demand fundamentals, or whether it’s coming from indiscriminate, cross-sector de-risking triggered by the narrative that “AI beats software.”
That was Arya’s interpretation, and I also don’t really understand this selloff.
If I had to find a reason, you could argue that there was simply too much leverage in the system recently, and mechanical selling—like CTA-driven flows—may have contributed.
Still, I think the magnitude of the drop is excessive.
I’m really bummed that I don’t have any cash right now, so I can’t buy more even after this drop.
tweet
Offshore
Photo
Michael Fritzell (Asian Century Stocks)
If M15X and P4 ramp up in February, then that's got to have an impact on prices, should it not
tweet
If M15X and P4 ramp up in February, then that's got to have an impact on prices, should it not
The increase in DDR5 prices at German retailers has come to an complete halt in recent days. Although many prices were still fluctuating, there was only an average price increase of +0.1% compared to mid-January.
https://t.co/cVLGc32iwW https://t.co/lWipvv4pOX tweet