Is this

More good news? 😃 https://www.reuters.com/world/asia-pacific/sk-hynix-talks-with-intel-about-deal-make-memory-chips-us-first-time-sources-say-2026-09-16/

AI slow down narrative and why we are taking a hit

I’ve looked into this extensively, and the whole situation seriously stinks. First, the security incident involving Hugging Face appears to have been driven, at least in significant part, by human prompting and interaction. That alone should make us extremely cautious about presenting the incident as evidence that AI systems are suddenly autonomous, uncontrollable, or on the verge of AGI. And, importantly, open-source models and developers were involved in identifying and stopping the problem. That part of the story seems to get far less attention. To me, the bigger issue is the narrative being constructed around incidents like this: “Look how dangerous AI is. Look how close we are to AGI. We need urgent regulation.” I don’t buy that argument. AGI is still a long way off. Current AI systems don’t think, feel, or understand the world in the way humans do. They can perform astonishingly sophisticated tasks, but that doesn’t mean they possess consciousness, intentions, or human-like reasoning. Even many of the researchers working at the frontier understand the distinction. At the same time, they understandably support some form of regulation because they know these systems will become increasingly consequential. But there is another issue underneath all of this: open source is coming. It is only a matter of time before open-source models catch up with, and in some areas surpass, proprietary systems. And if highly capable AI becomes decentralized and widely available, that fundamentally threatens the business model of companies trying to build closed AI ecosystems. That is where my concern lies. If the US introduces AI regulations that effectively make frontier AI accessible only to a handful of enormous companies, the rest of the Western world is likely to follow. Other countries may adopt similar frameworks. The result could be a regulatory environment that doesn’t simply make AI safer—it could make closed, centralized AI the default and open-source AI increasingly difficult to develop. Whether that is intentional or not, the effect could be the same. And I think that would be a huge mistake. Trying to suppress open-source AI will not necessarily make AI safer. It could actually make the ecosystem less transparent, less accountable, and potentially more dangerous. When development is concentrated inside a handful of corporations, the public has far less ability to inspect, test, challenge, reproduce, or improve what those systems are doing. That is the real danger: the black box. Look at the open-source projects that underpin much of the internet. When something goes wrong, thousands of developers can inspect the code, identify vulnerabilities, propose fixes, and independently verify solutions. Open source has repeatedly been an important part of finding and fixing catastrophic vulnerabilities precisely because it allows many people to see and work on the same problem. The same principle could become incredibly important for AI. Open-source AI could be particularly valuable in medicine because individual doctors, researchers, hospitals, and specialists could contribute knowledge and improvements rather than relying entirely on a single corporation’s model. The same applies to law, science, engineering, education, and countless other fields. Of course, open-source AI needs safeguards. Nobody should pretend there are no risks. But the answer cannot simply be to put the most powerful AI systems behind the walls of a few companies and assume that those companies will somehow be better stewards of humanity’s future. And then there is China. China is going to continue developing increasingly capable models, and those models will inevitably become attractive to companies and countries that want powerful AI without depending entirely on Western corporations. China itself also has reasons to be wary of uncontrolled open-source development because decentralized technology is inherently harder for governments to control. But developers are not going to stop. The technology will spread. Trying to prevent that through regulation may therefore produce the opposite of what policymakers intend. Instead of eliminating the risk, it could create two parallel ecosystems: tightly controlled corporate AI on one side and increasingly sophisticated decentralized models on the other. And the decentralized ecosystem may become harder to control precisely because people are trying to control it. That is why I think the current conversation needs to be much more intellectually honest. If the argument is genuinely about safety, then let’s talk about safety. If the argument is about preventing dangerous capabilities from becoming widely available, then let’s have that debate. But if the underlying objective is to protect the market position of a handful of companies by making it prohibitively difficult for open-source developers to compete, then we should call that what it is. AI should not become another technology where a handful of corporations control the infrastructure, the models, the information, and ultimately the decisions that affect everyone else. The idea that society should voluntarily become dependent on a few private companies for intelligence is absurd. They would be incredibly short-sighted if they believed people would simply accept that forever. Open source isn’t going away. Trying to kill it may not make AI safer. It may simply make the eventual decentralized AI ecosystem more determined, more sophisticated, and more difficult to control.

Intel updated valuation 2027

(We are so undervalued Intel will pull a Micron or a Nvidia) 2027 market will wake up to Intel growth. We are so undervalued that I revised my numbers down to be conservative but I believe that my numbers below are achievable. 1T valuation next year for Intel is more than achievable based on the revenue forecast and continued IFS success… My Consolidated Intel Revenue & Valuation Model Based on my consolidated framework, I project Intel's revenue could nearly double in 2027, assuming current pricing gains remain sustainable and TAM expansion holds at ~20%. 2027 Revenue Outlook by Business Segment * Data Center & AI (DCAI) | $35–40B: I expect Intel to exit 2027 at a run-rate substantially higher than current consensus expectations, driven by server CPU share gains, favorable pricing dynamics, and sustained demand across data center and AI infrastructure. * Client Computing (CCG/CCPG) | ~$40B: I project incremental market-share capture from AMD as Intel’s product availability and manufacturing capacity improve, allowing the company to satisfy demand and capitalize on favorable ASP trends. * Intel Foundry (IFS) | ~$31B: While my underlying model indicates potential for ~2× YoY growth, I have moderated my working assumption to ~$31B for conservatism rather than using the full upside case. Manufacturing & Capacity Utilization A core driver in my model is the normalization and expansion of Intel’s internal wafer supply. Scaling production capacity across Intel 3, 7, 18A, 18A-P, and 14A nodes will allow a greater proportion of product demand to be manufactured internally, expanding available wafer capacity and eliminating supply constraints. Model Sensitivities & Execution Risks I treat this near-doubling outcome as a scenario dependent on successful execution across both the product and manufacturing roadmaps—not as a base-case forecast. The primary sensitivities in my model are: * ASP durability and TAM expansion rate (~20%) * Competitive share capture (particularly vs. AMD) * Wafer availability and the execution pace of advanced node ramps Valuation Trajectory * 2027 Target ($2.1T): If my model's execution assumptions hold, my valuation table points to Intel reaching a $2.1T market cap next year. (This is based on price to sales as Capex will significantly increase) assuming a stable increase in profitability Intel will easily reach the 1T valuation * 2028 Projection ($4T): Extrapolating these assumptions into 2028 yields a $4T market cap. However, I do not assume this same revenue expansion rate can simply be carried forward, as the emergence of physical AI will fundamentally alter market expectations and my underlying assumptions.

Intel Brief: Recent Highlights & What’s Next

​Intel completed a massive equity raise to fund its expanding U.S. manufacturing footprint and deepened its partnership with ASML to scale High-NA EUV lithography for future nodes. ​Data Center and AI revenue jumped significantly year-over-year. Enterprise cloud demand for Xeon 6 server CPUs has been strong as companies build out heavy AI infrastructure. ​Aggressive pricing on Arrow Lake. Refresh processors allowed Intel to temporarily capture over 21% of Amazon US desktop CPU sales, though discounts are tapering off to protect margins. ​Monthly yields on the key Intel 18A process node continue to improve, easing production bottleneck concerns. ​Forward-Looking Outlook ​Panther Lake for laptops and Wildcat Lake for Edge/Series 3 are ramping up on Intel 18A, with over 200 client design wins locked in. ​Intel Foundry is aggressively pitching itself as a viable alternative to TSMC, securing early customer interest for its upcoming 14A node and advanced packaging services. ​Wall Street analysts have steadily revised 2026/2027 earnings expectations upward, backed by gross margins recovering toward 41%. ​Supply chain bottlenecks and persistent weakness in the broad consumer PC market remain active headwinds. ​TL;DR: Execution is improving, the server market is providing solid momentum, and 18A manufacturing is making real progress—making Intel a key turn-around story to watch.

So...

Back to 98 if oil keeps rising