Nusantara Capital

You guys shouldn’t defend the Nicko case. It’s not the same as Nadiem’s situation.

We are making our discount permanent! 🎉 Enjoy building with DeepSeek-V4-Pro and bring your innovative ideas to life! 🚀

Gak ada yang nanya ke TVRI kenapa hak siar Piala Dunia untuk OTT dikasih ke pihak yang bahkan sebelumnya ga punya OTT?

Maaf gaes aku kalap... Langsung buka rekening USD & SGD pas dollar tembus 17.750an 🫣🫣🫣

you guys really have to gratefull kedepannya bakal lebih WOW

Asli. gw maen twitter dari 2012 dan gak pernah timeline gw segini tense dan ributnya😂😂😂

Mulai ngumpulin cash dlm bentuk HKD. Kita gapernah tau cobaan apa yg ada di depan. Maka dr itu ada baiknya punya worst case scenario. Just in case..

The bull case for Cerebras ($CBRS) is fundamentally a view that the AI infrastructure stack is entering its post-GPU monoculture phase. For years, Nvidia dominated because the market needed a general-purpose AI compute standard quickly. But as AI workloads scale exponentially, inference increasingly becomes the real bottleneck rather than just training. That is where Cerebras becomes interesting because its wafer-scale architecture is not simply another chip, but an entirely different compute philosophy. Instead of stitching together thousands of smaller GPUs across increasingly complex interconnects, Cerebras built an entire wafer-sized processor optimized for ultra-low latency, massive memory bandwidth, and simplified scaling. In a world where AI models are becoming larger, more reasoning-intensive, and increasingly real-time, reducing communication overhead becomes strategically important. The market is slowly realizing that inference economics may ultimately matter more than benchmark leadership. AI monetization at global scale requires lower latency, lower energy intensity, faster token generation, and ultimately lower inference cost. That is precisely the segment Cerebras is targeting. The OpenAI relationship matters because it validates that frontier AI labs themselves are actively exploring alternatives to traditional GPU-heavy architectures. Once hyperscalers and frontier labs begin diversifying compute supply chains, the addressable opportunity expands dramatically. This is also not purely a semiconductor story. It is an infrastructure scarcity story. The world is simultaneously trying to build AI agents, reasoning models, robotics, autonomous systems, sovereign AI stacks, and hyperscale inference networks. All of these require exponentially more compute while the global AI supply chain remains constrained by advanced packaging, networking bottlenecks, and power availability. Importantly, markets increasingly reward companies perceived as owning AI toll roads rather than downstream applications because infrastructure captures economic rent while applications eventually face commoditization pressure. If Nvidia became the first great AI infrastructure winner of this cycle, investors are now searching aggressively for the next strategic infrastructure layer. Of course, the risk is obvious. Cerebras currently trades at extremely aggressive valuations, approaching roughly 200x sales, which means investors are already pricing in extraordinary future success long before financial results fully justify it. The probability of overpaying is very real, especially in a market environment where AI enthusiasm can detach prices from near-term fundamentals. But that is also the beautiful part of US capital markets. American markets have historically been willing to fund founders with ambitious visions long before the economics become fully visible. That willingness to aggressively finance technological ambition is precisely why so many transformative companies emerge from the US ecosystem in the first place. Investors are not simply valuing present cash flows. They are underwriting the possibility that a company becomes foundational infrastructure for the next technological era. Cerebras does not necessarily need to defeat Nvidia outright for the equity story to work. It simply needs to become strategically indispensable within the next generation of AI inference infrastructure. In a world moving toward exponentially larger AI workloads, even partial success could still create an enormous company.

这次英伟达老板黄仁勋，没有跟随特朗普 直接跑到胡同去吃方砖厂炸酱面了 该炸酱面荣获多年的米其林推荐

Kalo pemerintah beneran kerja, itu sebenernya bisa loh. Contohnya Mayor NYC Zohran Mamdani waktu awal menjabat nerima warisan defisit $12 miliar. Sekarang defisitnya udah berkurang jadi NOL (setelah 5 bulan menjabat) dengan cara: 1. Dapetin tambahan bantuan negara dari Albany + restrukturisasi pensiun 2. Tax the rich! Pajak baru pied-à-terre buat rumah mewah kedua ($500 juta/tahun) + penghematan operasional kota $1,77 miliar Tanpa naikin pajak secara besar-besaran buat rakyat biasa, malah tetap bisa nambah dana buat taman, perpustakaan, perumahan publik, childcare, dll. Jadi selama ini rakyat emang sering dikadalin ama Dajjal aja wkwk 😂

Khamzat is prob gona win but this is a closer fight than people think.. Maybe dont bet the house on strickland but a few g just for fun would be interesting..