AI Weekly Malaysia

Nvidia has doubled the MSRP of the RTX PRO 6000 Blackwell to $16,000, up from sub-$8,000 pre-order pricing last year. The card features 96GB of VRAM, making it a key option for local LLM inference and fine-tuning workloads.

Why: If you were budgeting for local GPU hardware to run large models, your cost just doubled overnight — recalculate build-vs-cloud-rental math now. For Malaysian builders importing GPUs, the ringgit impact is even steeper given currency conversion on top of the doubled USD price.

CoreWeave's Q2 2026 revenue doubled YoY to $2.575B, but operating expenses of $2.624B produced a $49M operating loss and $626M net loss, with total debt at $35.6B. 93% of revenue growth came from existing customers, and just three customers accounted for 72% of quarterly revenue. CEO Michael Intrator pitched AI compute as a continuous recurring loop (training, inference, evaluation, redeployment) rather than a one-time training cost, with managed inference services targeting $250M ARR by end of 2026.

Why: If you rent GPU capacity from neoclouds like CoreWeave, this signals pricing and service-model shifts ahead: they are pushing up-stack into managed inference and are financially stretched enough that contract terms or availability could change. The extreme customer concentration (three clients = 72% of revenue) and $35.6B debt mean builders should avoid single-provider lock-in for critical inference workloads and evaluate whether the 'continuous compute loop' framing matches their actual usage pattern before committing to long-term contracts.

CoreWeave CEO Mike Intrator says the company has signed A100 GPU contracts extending into 2029, demonstrating that Nvidia's 2020-era GPUs remain profitable nine years post-deployment. Power constraints and legacy infrastructure costs make older GPUs economically viable even as newer chips arrive.

Why: If you're budgeting GPU compute for AI workloads, don't assume older GPUs like the A100 will become cheap or obsolete soon — CoreWeave is locking customers into multi-year A100 contracts through 2029, which signals sustained pricing power for legacy hardware. This affects cost planning for anyone renting cloud GPU capacity or deciding whether to wait for next-gen capacity versus contracting now.

Nvidia has lined up $500 billion in financing through agreements with six major Wall Street firms (BlackRock, Blackstone, Apollo, KKR, Brookfield, Goldman Sachs) to fund AI infrastructure buildout, treating chips as long-term financial assets. Analysts warn that if China floods the market with low-cost compute, rapid hardware depreciation could crash the collateral values backing these loans, pushing investor yield demands to 11-17%.

Why: If Chinese low-cost compute enters the market and accelerates GPU depreciation, cloud compute prices could drop significantly — builders and founders should factor in the possibility of much cheaper inference costs within 1-2 years when making infrastructure and pricing decisions, rather than locking into long-term GPU commitments at today's rates.

The trycua/cua project published a blog post on using GPU passthrough in macOS VMs to accelerate llama.cpp-based LLM inference on Apple Silicon. The post details the technical approach for passing Apple GPU resources through to a virtualized macOS environment.

Why: If you run local LLM inference on Apple Silicon but need VM isolation for CI, agent sandboxes, or multi-tenant setups, this approach could let you keep near-native Metal GPU performance inside a VM rather than falling back to CPU-only inference. Builders evaluating local agent infrastructure should test whether the passthrough overhead is acceptable for their workload before committing to a bare-metal-only deployment.

GeForce RTX 50-series GPU prices have spiked sharply in the US, with Newegg median listings showing the RTX 5070 up 36% and the RTX 5060 up 27%, with some cards rising as much as 39%. The article frames this as Blackwell price hikes hitting the US market.

Why: If you budget for local AI/ML workloads or fine-tuning on consumer GPUs, these US price spikes likely signal similar upward pressure in Malaysia through import and distribution channels. Anyone planning a workstation build or GPU upgrade in the next quarter should lock in pricing now or reconsider whether cloud GPU rental (e.g., RunPod, Lambda, or local cloud credits) is cheaper than buying at these inflated levels.

CoreWeave reported Q2 revenue of $2.6B, up 112% YoY from $1.2B, with Q3 guidance of $3.4-3.6B, driven by hyperscaler demand for AI compute capacity. The company remains unprofitable—operating expenses more than doubled and marginally exceeded revenue. Nebius also surged 34% in a broader neocloud rally.

Why: CoreWeave's doubling revenue and continued losses signal that GPU compute demand is still outpacing supply economics—meaning builders shipping AI workloads should expect sustained or rising compute costs and should lock in capacity or explore alternative providers like Nebius before pricing tightens further. The fact that operating expenses exceed revenue at this scale suggests neocloud pricing power is not yet translating to margins, which could drive future price hikes.

CoreWeave reported Q2 2026 revenue of $2.58 billion (up 112% YoY), beating consensus by a narrow margin, while net loss widened to $626 million from $290 million a year prior. The GPU cloud provider carries $35 billion in debt against a $104 billion revenue backlog, with 1.5 gigawatts of active power, and announced a $21 billion Meta deal plus a multi-year Anthropic agreement in the quarter.

Why: If you're budgeting AI compute costs, CoreWeave's $35B debt load and widening losses signal that GPU cloud pricing is subsidized by aggressive capital expenditure that may not be sustainable long-term—consider locking in longer-term contracts or diversifying across providers (AWS, Google, Azure, CoreWeave) before pricing dynamics shift. The $104B backlog also indicates GPU capacity remains heavily pre-committed by large labs, which could squeeze availability and pricing for smaller builders.

Modular's Mojo programming language reached its 1.0 milestone, offering a Python-like syntax with Rust-like memory safety designed to unify AI workloads across GPUs, CPUs, and ASICs without vendor lock-in to CUDA or ROCm. Chris Lattner (creator of LLVM, Swift, MLIR) leads the project; Modular was acquired by Qualcomm in June 2026. The standard library ships under Apache 2.0 with LLVM exceptions, but the compiler itself is not yet open source—Modular says that may happen at Modcon next week.

Why: Mojo 1.0 stabilizes the language surface, but the compiler remains closed and Qualcomm's acquisition creates real uncertainty about governance and hardware neutrality. If you're evaluating alternatives to CUDA for AI inference, wait for the compiler open-sourcing before committing—Lattner's team says it could land at Modcon, but until then you're betting on a Qualcomm-owned stack. The MAX inference framework pairing is the practical entry point if you want to experiment today.

Nvidia secured commitments from Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR for up to $500B to build AI data centers, with Nvidia guaranteeing that GPUs used as collateral retain their value—covering up to 25% of any shortfall if liquidated chips fetch less than book value. The plan aims to create a secondary market for aging GPUs so demand persists as hardware ages, but creates 'wrong way' risk where Nvidia's obligations grow precisely when demand weakens.

Why: If a used-GPU market materializes, GPU compute prices could eventually drop for builders who rent capacity from neoclouds or data centers—relevant to Malaysian startups running inference workloads on cloud GPU services. But the more immediate signal is that Nvidia is financially engineering demand for its own chips, which means current GPU pricing power stays with Nvidia for now; don't plan infrastructure budgets assuming cheaper compute is coming soon.

GPU cloud provider Nebius plans to bring online over 1 GW of datacenter capacity annually starting 2027, funded by $20-25B in 2026 capex, $9B+ in customer prepayments, and asset-backed debt using GPUs as collateral. The company projects $20-25M revenue per MW for medium-term leases and $40-50M for short-term leases under six months.

Why: If you rent GPU capacity for AI workloads, Nebius's aggressive buildout signals more supply coming online by 2027, which could ease pricing pressure—but the heavy reliance on customer prepayments means large tenants are locking in capacity now, potentially squeezing spot or short-term availability for smaller builders. The $40-50M/MW short-term lease revenue figure tells you GPU cloud margins on urgent demand remain extremely high, so negotiate early and long if you have predictable workloads.

Nvidia is reportedly testing reduced memory configurations for its upcoming Rubin Ultra AI accelerator due to HBM supply shortages, with designs including as little as 192 GB and a step back to HBM4 from a more advanced memory type. The report is based on supply-chain rumors, not official confirmation.

Why: If Rubin Ultra ships with less memory than originally planned, AI/ML teams building large-model inference or training pipelines should factor tighter VRAM ceilings into their 2026-2027 infrastructure roadmaps — especially in SEA where GPU access is already constrained by allocation priority. Founders budgeting for next-gen GPU rentals or cloud instances should not assume memory specs will scale up linearly from current Blackwell-class hardware.

A GeForce NOW exploit reportedly allows users to access the full Windows desktop through a simple file swap, bypassing NVIDIA's sandbox restrictions. A modder used this to run local AI models on the Ultimate tier, which provides 48GB of VRAM. The article body itself contains only website boilerplate with no additional technical detail.

Why: If this exploit persists, it effectively turns a ~RM100/month cloud gaming subscription into a 48GB VRAM GPU rental for AI inference, which is dramatically cheaper than equivalent cloud GPU instances. Builders experimenting with large local models should note this exists but expect NVIDIA to patch it quickly and enforce ToS violations.

Analyst Claus Aasholm estimates that Amazon, Alphabet, Meta, and Microsoft collectively hold nearly $2 trillion in purchase commitments for AI hardware and memory as of Q2 2026, with Alphabet leading at $811 billion and Apple trailing at $57 billion. A significant portion targets memory components, reflecting a shift from Apple's historical dominance in long-term component contracts to hyperscalers driving the market.

Why: If you're budgeting for GPU or AI inference costs over the next 1-2 years, this signals sustained pricing pressure and scarcity for AI hardware and memory — hyperscalers are locking up supply years ahead. Malaysian founders and developers relying on cloud AI compute should expect continued high costs for GPU-backed services and may need to weigh smaller-model or CPU-based inference strategies sooner rather than later.

French startup Kog, founded solo by Gaël Delalleau, claims 30x faster LLM inference on conventional datacenter GPUs (AMD MI300X, NVIDIA H200) via software optimization. Its demo hit 3,000 tokens/second but only with a 2B-parameter model (Laneformer 2B, now open-sourced), and Kog admits customers won't fine-tune small models, so it is pivoting to accelerate larger models — a claim still unproven. The startup generated 200 business leads and is targeting software engineering workflows where Claude Code users wait hours for results.

Why: The 3,000 TPS demo is real but narrow — it runs on a 2B model, not the large models production teams actually use. Builders should treat the '30x faster' headline as aspirational until Kog shows results on production-scale models. The open-sourced Laneformer 2B is worth examining if you work on inference optimization, but don't change your serving stack based on this.

Nvidia is partnering with Apollo Global Management, Blackstone, BlackRock's Global Infrastructure Partners, Brookfield Asset Management, Goldman Sachs, and KKR on a $500 billion AI infrastructure financing package, with an announcement expected as early as Monday. The deal signals private capital is now a primary funding source for AI compute buildout.

Why: If this $500B materializes, it means GPU compute supply will scale aggressively over the coming years, which could eventually ease pricing pressure for builders running inference workloads. For SaaS founders and AI/ML practitioners, this is a signal to avoid over-committing to long-term GPU lock-in contracts right now, since cheaper or more available compute may follow the infrastructure buildout. No immediate action required—this is a financing arrangement, not a product or API change.

VectorWare demonstrates that Rust's portable SIMD (core::simd) can now target GPUs by mapping a Simd<T, N> vector directly onto a GPU warp's 32 lanes—for example, Simd<i16, 32> assigns one i16 element per lane, and vector addition compiles to a single warp instruction. This builds on their earlier work bringing std::thread to GPUs, completing a parallelism hierarchy where CPU threads contain SIMD lanes and GPU threads (warps) serve the same role.

Why: If you write Rust for performance-critical workloads, this shows that core::simd abstractions can now span both CPU and GPU targets without architecture-specific intrinsics—meaning one codebase could potentially target x86, Arm, and NVIDIA GPUs. However, this is a VectorWare product announcement with no benchmarks, pricing, or availability details, so there is nothing concrete to adopt or change today.

AMD's Radeon RX 9070 GRE with 12GB of VRAM is back at $499, reportedly the only current-generation graphics card selling below its MSRP. The article is primarily a deal alert with no benchmark or technical analysis included in the available text.

Why: For builders running local AI inference or fine-tuning, 12GB VRAM at $499 (~RM2,200+ before local markup) is a viable entry point for sub-13B parameter models, but Malaysian buyers should expect import duties and retailer margins to push the real landed price higher. Decide whether to wait for local availability or import directly based on your actual VRAM budget needs.

CNBC's Morning Squawk newsletter covers August 12 2026 market movers: July CPI expectations (0.1% monthly, 3.4% annual), CoreWeave's revenue surge, election betting markets, JPMorgan's 2028 Olympics investment, and rising oil prices amid U.S.-Iran uncertainty. The piece is a broad pre-market roundup with no technical depth on any single topic.

Why: CoreWeave's revenue surge signals continued tight demand for GPU cloud infrastructure, which affects pricing and availability for AI/ML builders and startups relying on rented compute. However, the article provides no figures or detail on CoreWeave's numbers, so there is no actionable takeaway beyond a general signal that GPU cloud capacity remains a seller's market.

Tom's Hardware highlights a Walmart deal on the 15.6-inch MSI Cyborg 15 laptop with an RTX 5060 GPU, priced at $949 after a $350 discount. The article is primarily a retail deal listing with no technical benchmarks or hands-on review content.

Why: This is a US-only consumer gaming laptop deal with no direct relevance to Malaysian builders; the RTX 5060 may be worth tracking for budget local AI/ML experimentation, but no actionable decision can be made from this listing alone.

CNBC's Jim Cramer highlighted benign July CPI data and strong earnings from AI cloud providers CoreWeave (shares +19%) and Nebius (shares +27%), driven by massive demand for AI cloud computing. The piece is investor commentary tying macro inflation data to AI infrastructure stock performance, with mentions of Nvidia, Intel, and Micron as related plays.

Why: For builders choosing GPU cloud providers, CoreWeave and Nebius posting strong earnings signals sustained capacity expansion and competition in AI compute — but this article gives zero technical, pricing, or product detail to act on. It is stock market commentary, not infrastructure news. No actionable takeaway for shipping decisions.

A modder attached a Radeon RX 9070 XT as an external GPU to a Steam Machine, achieving over 100 FPS in Crimson Desert on High settings. The setup involved relocating the boot drive to a USB-C port and using an M.2-to-OCuLink adaptor with an eGPU dock.

Why: This is a niche hardware modding feat with little practical takeaway for builders shipping software or AI workloads. The OCuLink eGPU approach may be worth noting if you are exploring compact GPU compute setups, but there is no actionable decision here for most developers or founders.