Inference-Optimized Silicon

CoreWeave announced availability of the NVIDIA Vera Rubin NVL72, with Cognition as its first production customer, alongside new support for the NVIDIA Vera CPU. Cognition, which uses CoreWeave for training, reinforcement learning and inference, reported that the Vera Rubin NVL72 delivered up to 4.8x higher total token throughput for SWE-2 inference workloads versus a GB200 NVL72 baseline and 3.8x higher output-token throughput for reinforcement-learning workloads. CoreWeave said a Vera rack can contain 128 CPUs and 11,264 cores, theoretically supporting more than 11,000 concurrent isolated environments, and that testing showed more than three times faster agent sandbox startup times compared with an x86 CPU. The company will offer Vera on bare metal using the same operating model and economics as the rest of its infrastructure, aiming to monetize CPU-intensive infrastructure alongside accelerator hours. CoreWeave remains heavily dependent on NVIDIA's technology roadmap and faces competition from hyperscalers and specialized GPU clouds including Microsoft Azure and Nebius Group N.V., which closed four deals in the quarter averaging more than $1 billion each and plans roughly £1.7 billion in U.K. AI compute expansion expected to deliver 65 MW when fully operational in 2027.

Volantis, a semiconductor company building a new category of AI inference system, announced an $88 million Series A co-led by Lachy Groom and Abstract Ventures, with participation from John Doerr, VXI Capital, Triatomic and Susa Ventures, plus angel investors Dwarkesh Patel, Naveen Rao and Sholto Douglas. The company's first system, A-1, is being designed to run models exceeding 20 trillion parameters at up to 10,000 tokens per second per user while reducing inference cost per token. Volantis says A-1 will increase memory capacity and bandwidth simultaneously by nearly two orders of magnitude, using a photonic interconnect that connects compute chips to memory and aggregates the bandwidth of large numbers of memory chips into a unified pool. The architecture uses custom micro-VCSELs rather than external lasers, drawing on the existing gallium arsenide VCSEL supply chain and avoiding indium phosphide supply constraints, with end-to-end links consuming less than one picojoule per bit. The financing will support development and commercialization of A-1 and its photonic memory architecture, including expanding the engineering team, and Volantis plans to deliver its first integrated inference engines to customers in 2027.

Cerebras Systems is carrying $25.4 billion in remaining performance obligations as of June 30, 2026, work customers have signed for but the company has not yet delivered. That backlog sits alongside management's raised August forecast for core revenue of $880 million to $890 million for all of 2026, and a 2027 target for core revenue to more than triple. Core revenue in the second quarter of 2026 was $209.9 million, up 103% from a year earlier, most of it tied to the company's cloud service as its OpenAI deployment ramped up. The signed work includes no business from AWS or any other hyperscaler, though Cerebras expects its offering on AWS' Bedrock platform to be generally available in the first quarter of 2027. Management has secured more than 600 megawatts of data center capacity, live now or due by the end of 2027, and said data center space is the industry's bottleneck; core gross margin should hit its low point in the third quarter of 2026 before a significant fourth-quarter improvement. The shares trade 37.3% below their 52-week high at 51.1 times sales, against 3.1 times for the S&P 500.

Lenovo announced Lenovo AI Express, a new quick-start program developed with NVIDIA that accelerates the path to production AI by cutting order-to-ship times to as few as 15 business days for eligible configurations. The program offers three validated quick-start configurations of the Lenovo Hybrid AI Factory with NVIDIA: a Small tier shipping from 15 days for inferencing for tens of users at 30+ TPS with model sizes from 7B to 70B, built on Lenovo ThinkSystem SR650a V4 with two NVIDIA RTX 6000 PRO Blackwell Server Edition GPUs; a Medium tier from 20 days for hundreds of users with model sizes from 70B to 400B, built on Lenovo ThinkSystem SR675 V3 with eight NVIDIA RTX 6000 PRO Blackwell Server Edition GPUs; and a Large tier from 25 days for thousands of users with up to a trillion parameters, built on Lenovo ThinkSystem SR680a V4 with NVIDIA HGX B300. Lenovo cited its 2026 CIO Playbook finding that organizations expect an average $2.79 return for every $1 invested in scaled AI execution, with 93% of enterprise respondents anticipating positive returns. Ashley Gorakhpurwalla, President of Lenovo's Infrastructure Solutions Group, said the program gives customers an accelerated path to deploying AI infrastructure while mitigating the risk of overbuilding or costly delays, and NVIDIA vice president of enterprise platforms Chris Marriott said the ready-to-ship solutions combine Lenovo's validated systems with NVIDIA accelerated computing, networking and AI software. The configurations support the latest AMD and Intel CPUs and can be extended with Red Hat AI Factory with NVIDIA or NVIDIA AI Enterprise software, plus Veeam Kasten for AI data resilience, and Lenovo is expanding access through its global partner ecosystem via the Lenovo 360 framework.

SiMa.ai, a startup developing chips and software that let robots, drones, cameras, and other devices run AI directly on the device, has raised a $150 million Series C at a $1.45 billion valuation. The round was co-led by Fidelity Management & Research Company and Amplify, with participation from Alter Venture Partners, Dell Technologies Capital, and StepStone Group. Founded in 2018 by Krishna Rangasayee, previously COO of chipmaker Groq, the company provides energy-efficient chips that eliminate the need to send data back and forth to the cloud. SiMa.ai hopes its low-latency performance and more affordable chips, compared to Nvidia's GPUs, will help it capture the growing market for physical AI devices, including humanoid robots. The new round brings SiMa.ai's total capital raised to over $500 million, and the startup was previously valued at $960 million after raising an $85 million Series B in July 2025, according to PitchBook.

Advanced Micro Devices announced a collaboration with Advantech to accelerate Edge AI workloads using AMD embedded and GPU platforms, while MindWalk has moved OpenFold3 into production on AMD Instinct GPUs. The Advantech alliance focuses on scalable, partner-led Edge AI solutions, with Advantech's WEDA ecosystem putting AMD Ryzen AI Embedded processors into pre-validated edge platforms that can be rolled out across industrial fleets. MindWalk's OpenFold3 work on Instinct MI325X GPUs reaches into life sciences prediction workloads, supported by inference microservices for regulated, production-grade scientific use cases. The company said the announcements line up with its existing narrative of expanding AI infrastructure across accelerators, software and full systems, and that MindWalk's validation of AMD Inference Microservices and the Kubernetes native path into regulated production supports the view that ROCm and related tooling can lower switching costs for third party models. AMD investors are watching how many additional production case studies using AMD Instinct GPUs and inference microservices appear through 2027 across sectors like life sciences, security and edge industrial systems.

The global Vision-Language Models market is projected to expand from approximately USD 3.84 billion in 2025 to USD 41.75 billion by 2035, registering a compound annual growth rate of 26.95% between 2026 and 2035, according to a report added to ResearchAndMarkets.com's offering. Growth is being driven by rising enterprise demand for multimodal artificial intelligence, rapid advances in computing infrastructure, and the integration of visual reasoning into industrial and domain-specific workflows. Hyperscale hardware platforms, including NVIDIA Blackwell GPUs and the Cerebras Wafer-Scale Engine 3, are providing the processing capacity required to train and deploy increasingly sophisticated VLM systems. By model type, image-text Vision-Language Models accounted for the largest share at 44.5%, while cloud-based deployment represented 66% of total market revenue and IT and telecom led by industry with a 16% share. North America held 45% of global Vision-Language Models market revenue in 2025, supported by strong model development capabilities, extensive cloud and data center infrastructure, and early enterprise adoption of reasoning-focused architectures such as Gemini 2.5 Pro and GPT-4.1. Object hallucination remains a significant barrier to large-scale adoption, with leading models continuing to record an industry-standard error rate of approximately 3%.

CoreWeave has become the only cloud provider to earn SemiAnalysis' Platinum ClusterMAX rating in all three ClusterMAX evaluations to date, the company announced. The SemiAnalysis ClusterMAX Rating System is an independent benchmark for how cloud providers handle large-scale AI workloads; in its latest report, SemiAnalysis found CoreWeave's health checks worked as intended with excellent reliability and that nearly all tests reached expected values out of the box. When the first ClusterMAX rating was published in early 2025, it evaluated roughly two dozen providers and found exactly one Platinum, CoreWeave; when ClusterMAX 2.0 followed later that year, the field had more than tripled to 84 providers and CoreWeave remained the only one at the top. CoreWeave co-founder and chief technology officer Peter Salanki said three straight independent assessments, each against a higher bar and a more crowded field, matter to customers deciding where to run workloads they cannot afford to get wrong, while SemiAnalysis founder and chief executive Dylan Patel called CoreWeave the operational benchmark for the industry. CoreWeave also cited record-breaking MLPerf benchmark results and a number one ranking for inference speed and price-performance for Moonshot AI's Kimi K2.6 in independent inference benchmarking by Artificial Analysis.

Apple is developing enterprise AI servers built around future M8 Ultra processors, with designs under consideration using either two or four of the chips and a possible 2029 launch aimed primarily at AI inference, The Information reported on September 16. Apple has also considered using Nvidia's NVLink Fusion technology to connect those processors, though neither the server nor Nvidia's involvement is finalized and the project could still change or disappear entirely. For Nvidia, NVLink Fusion extends its interconnect strategy to custom CPUs and accelerators made by other companies, creating a second line of defense around its AI franchise; networking equipment can represent 10% to 15% of AI data-center hardware costs, giving Nvidia a meaningful revenue pool beyond GPUs even if it loses higher-value compute sales. Apple's M-series chips already combine performance with relatively efficient power consumption, and selling complete servers would let it pursue AI developers, governments and businesses that want local inference, but a 2029 target leaves years for competitors to improve and would mark a re-entry into a server business Apple abandoned when Xserve disappeared in 2011. Insider Monkey tracked 169 hedge funds holding Apple in the second quarter versus 170 in the first, with Arrowstreet Capital increasing its position 54% to roughly 29.9 million shares, while Nvidia rose to 285 hedge-fund holders from 275 as Arrowstreet increased its NVDA stake 10% to about 34.7 million shares.

Intel's CEO said the company's current CPU production is meeting only about 50% of customer demand for AI workloads, signaling tight supply as AI-related order interest strengthens. Intel and rival CPU suppliers reported stronger AI-related order interest after Meta's Muse agent news in September 2026, with chipmakers tied to Muse-driven AI compute needs flagging rising pressure on CPU capacity planning across upcoming production cycles. Meta's Muse agents are pulling more AI inference onto general-purpose CPUs, which plays directly to Intel's existing Xeon and Core Ultra footprint, and the 50% fulfillment rate points to full factories and tight allocation rather than idle capacity. Intel designs and manufactures CPUs and related computing hardware across the US, Ireland, Israel, and other regions, so the squeeze in AI-focused processor demand intersects directly with its role as a large-scale producer in the global semiconductor industry. The most direct signal to track next is whether Intel starts disclosing materially higher Data Center and AI volumes tied to Muse-like agent workloads, alongside concrete updates on easing CPU backlogs.

Qualcomm shares jumped approximately 7% to $190.085 after the chip designer disclosed an AI-infrastructure partnership with Amazon that could see the e-commerce and cloud giant purchase as much as $60 billion of Qualcomm products. Under the arrangement, Amazon would receive purchase-linked warrants valued at roughly $4 billion, equivalent to about 6.7% of the potential $60 billion purchasing ceiling. The two companies are also building custom inference silicon and optical connectivity capable of reaching 1.6 terabits per second. The deal gives Qualcomm a credible path to becoming a larger supplier inside hyperscale AI infrastructure rather than remaining overwhelmingly tied to mobile chips. At $190.085, Qualcomm trades 8.3% above its $175.52 GF Value estimate, suggesting the market is already pricing in part of the Amazon opportunity.

Intel CEO Lip-Bu Tan said at Splunk's .conf26 conference in Denver this week that the chipmaker is meeting only about 50% of what its customers are asking for, as demand for its processors has outrun its factories. Tan attributed the shortfall to an explosion in demand for central processing units to run artificial intelligence inference, work that leans on CPUs rather than the graphics chips that dominate AI headlines. The shortage is visible in Intel's results: revenue growth accelerated from 3% in the third quarter of 2025 to 7% in the first quarter of 2026 and 25% in the second quarter, when revenue reached $16.1 billion, while the data center and AI segment grew 59% to $6.3 billion and the client computing and physical AI segment grew 13%. Intel guided third-quarter revenue to between $15.8 billion and $16.8 billion, implying growth of about 19% at the midpoint, and its non-GAAP gross margin climbed from 29.7% a year ago to 41% in the first quarter of 2026 and 41.8% in the second quarter, with management expecting 42% in the third quarter. Chief financial officer Dave Zinsner said the company is meaningfully increasing investments in equipment, clean room space, and substrates, and Tan said the 18A process behind its new Panther Lake laptop chips is in high-volume production while its successor, 14A, begins production in the first quarter of 2027.

Cerebras Systems announced on September 1 a new AI data center in Mikkeli, Finland, built with partner Compute Nordic Finland, that will grow in stages to 165 MW of contracted capacity, with construction on the first 50 MW already under way. The deal runs through a series of service orders, each with a seven-year term, stepping up from 50 MW to 80 MW and eventually 165 MW, and an independent study dated September 12, 2025 put the eventual regional investment at €1.0 billion to €1.7 billion. Cerebras reported on August 12 that cloud revenue for the quarter ended June 30 rose 281% from a year earlier, with $25.4 billion in customer obligations still to be delivered, while core gross margin reached 41%, roughly 940 basis points above a year earlier. On a GAAP basis, second-quarter gross margin was 14% and operating margin was negative 265%, and even the core measure that strips out stock compensation, warrant amortization and pass-through data center costs showed an operating margin of negative 16%. For the third quarter, Cerebras guided to core operating margin between negative 25% and negative 23% on core revenue of $214 million to $216 million, while hedge fund ownership rose from zero funds to 78 and short interest sits at 11.88% of the float.

Xeal launched Laitent, which it calls the world's first edge inference compute network using idle EV charging capacity, tapping more than 200MW of permitted, installed electrical infrastructure across 1,600+ properties. Xeal, a member of NVIDIA Inception, plans to deploy over 100,000 NVIDIA GPUs alongside EV charging infrastructure, and has secured partnerships with Rafay Systems for AI infrastructure orchestration, Spectrum Business for dedicated enterprise-grade fiber, dozens of real estate and property managers, and a Tier 1 inference provider for up to 5MW of compute. The first Laitent Pod will be brought online with partner JVM Realty by the end of 2026. Each Laitent Pod is about the size of one parking space, contains up to 48 NVIDIA Hopper or Blackwell Ultra GPUs, requires no water hookup, and runs quiet at less than 65 decibels, offering sub-20ms latency in metro areas. Xeal said EV charging sites typically operate at less than 10% of permitted capacity, and it taps the remaining 90% for compute, with property owners able to add as much as $1m in property value for little-to-no upfront investment. Looking beyond the initial 200MW of installed charging capacity, Xeal plans to unlock over 1GW of existing headroom across real estate and EV charging deployments.

Nvidia CEO Jensen Huang said the company expects chip sales next year to be about twice this year's level, sending shares up more than 2% Thursday. Speaking at an event in Scotland, Huang pointed to continued demand as businesses expand their use of artificial intelligence. Nvidia also released preliminary MLPerf results on Sept. 16 showing its next-generation Vera Rubin NVL72 platform delivering up to 3.7 times the inference throughput of the previous GB300 system on the Qwen3-VL test. Vera Rubin has entered full production, with shipments expected to begin this fall. The broader market added to the lift, as U.S. stocks rebounded Thursday with oil prices falling more than 2% and the 10-year Treasury yield easing, helping technology shares recover from recent pressure.

Bank of America estimates Meta Platforms could save roughly $8.5 billion in 2027 by running AI workloads on its own custom silicon instead of buying third-party chips, an outside analyst estimate rather than company guidance. The figure rests on a specific roadmap: Meta plans to deploy its third-generation MTIA 450 chip, code-named Arke, in the first half of 2027, followed by the higher-performance MTIA 500, or Astrid, later that year, both co-developed with Broadcom and aimed at AI inference workloads. BofA models Meta deploying 5 to 6 gigawatts of owned capacity in 2027 at a total cost of roughly $200 billion, assumes chips make up 60% of that spend, and pegs Meta's custom silicon as about 40% cheaper than third-party equivalents. Broadcom CEO Hock Tan said custom chips optimized for a customer's own workloads outperform any GPU and can do so at half the cost, and confirmed Broadcom will deliver three generations of MTIA accelerators to Meta between now and the end of 2027. Meta's FY2026 capex guidance sits at $130 billion to $145 billion, narrowed from $125 billion to $145 billion, with total expense guidance raised to $165 billion to $169 billion, while Q2 2026 revenue reached $60.80 billion, up 27.96% year over year, on advertising revenue of $59.36 billion.

Intel reported performance gains in its latest MLPerf Inference v6.1 results across Intel Xeon 6 processors and Intel Arc Pro B-series GPUs. Intel Xeon 6980P processors delivered 2.4x higher Llama 3.1 8B Server throughput and 56% higher Offline throughput than MLPerf v6.0 on the same hardware setup, while customer and partner results rose from 29 to 39, with Oracle, Red Hat, Quanta Cloud Technology and Supermicro making their first submissions. Intel expanded Xeon 6 participation from two to five processor models, lifting CPU inference results from 24 to 35, and its Arc Pro B70 GPUs supported workloads including Llama 2 70B, gpt-oss-120B and Whisper, with a four-GPU system offering 128GB of Video Random Access Memory and Server performance up 36% and Offline performance up 27% versus MLPerf v6.0. Intel also co-developed results for the new end-to-end retrieval-augmented generation benchmark, using a system that paired a Xeon 6787P processor with four Arc Pro B70 GPUs to split the workload between CPU and GPUs. Intel faces competition from Qualcomm, which is expanding into AI data-center infrastructure through its Dragonfly platform, and from AMD, which is strengthening its AI infrastructure with the Helios platform.

Meta Platforms is moving deeper into custom AI silicon, with a new generation of internally designed chips set to enter its data centers in the first half of 2027. The company is currently testing its third-generation MTIA 450 processor, code-named Arke, while its successor, MTIA 500, or Astrid, is expected to complete design work in about a month and reach data centers by the end of 2027. Meta is working with Broadcom on chip design and Taiwan Semiconductor Manufacturing Co. on production, and has committed to deploying more than a gigawatt of the chips over a 12-month period. Twelve Arke chips delivered by TSMC on Sept. 1 performed within 2% to 3% of Meta's simulations and have already run Meta models alongside models from DeepSeek and Alibaba. Meta also canceled its planned Olympus processor, which was intended to handle both AI training and inference, partly because of cost concerns, and is instead prioritizing inference, the day-to-day running of AI models.

Cirrascale Cloud Services announced the production release of the Cirrascale Inference Platform, a complete software stack for enterprise-grade AI inference, at the AI Infra Summit in Santa Clara. The platform lets enterprises run open-source models, their own private models, and closed model ecosystems from a single serverless platform, including Google Gemini delivered on premises through Google Distributed Cloud and operated by Cirrascale. Its model and hardware selection layer automatically routes each request to the right model and runs it on the best available accelerator across NVIDIA, AMD, Tenstorrent, and Qualcomm with no code changes required to switch, and teams can fine-tune models on their own private data without that data leaving their environment. The platform includes a turnkey private chat experience connected to a company knowledge base, built-in controls to manage AI spend across teams, and governance guardrails for agentic workloads, aligned with HIPAA, SOC 2, and FedRAMP requirements where required. The Cirrascale Inference Platform is available now across Cirrascale's U.S. and international regions.

Broadcom CEO Hock Tan said he sees little reason to change the company's longer-term AI outlook, pushing back on fears of a slowdown in frontier artificial intelligence development that sent chip stocks sharply lower on Monday. Speaking in a Monday CNBC interview, Tan said Broadcom still expects demand for computing infrastructure used for both AI model development and inference to remain durable, remarks that reinforced the company's existing forecasts for fiscal 2027 and 2028. The selloff followed comments from Anthropic CEO Dario Amodei calling for a more measured approach to advances in AI models, with OpenAI CEO Sam Altman and Elon Musk also raising concerns about the pace and risks of AI development. Tan expects Anthropic to become Broadcom's largest custom-chip customer in 2027 and to retain that position in 2028, overtaking Alphabet as the company's biggest custom-chip customer. He also pointed to inference, the running of trained AI models for everyday use, as another potential source of sustained demand.

AnalogAI has selected the memBrain Synaptic Analog Generative Engine neuromorphic hardware intellectual property from Microchip Technology's Silicon Storage Technology subsidiary for its first real-world edge AI processors. AnalogAI is using the SST memBrain SAGE IP to deliver analog compute-in-memory performance at or below one watt for ultra-low-power edge applications, targeting environment-adapting humanoid robots, drones and vehicles. Mark Reiten, senior vice president of Microchip's Intelligent Compute business unit, said the IP serves as the core inference engine for AnalogAI's first products, delivering the compute performance and power efficiency the company requires. AnalogAI chief executive officer Jaejun Lee said the company chose the silicon-proven memBrain SAGE IP after an industry-wide search because it accelerates development while meeting ultra-low-power and high-performance targets. The memBrain SAGE IP has been developed and deployed in 40 nm and 28 nm foundry processes using production-ready SuperFlash memory, with a roadmap that includes 22 nm development.

Dutch semiconductor startup Euclyd has raised $231 million in a Series A round backed by Samsung. The 200-million-euro round was co-led by Samsung alongside Somerset Capital Partners, the Scaleup Europe Fund managed by EQT, and Innovation Industries, according to CNBC. Founded in 2024 by Bernardo Kastrup and Atul Sinha at High Tech Campus Eindhoven in the Netherlands, Euclyd is building chip systems designed to reduce the energy consumption and cost of running AI models, targeting AI inference rather than model training. Chief executive officer Bernardo Kastrup told CNBC that Samsung brings more than capital, citing its memory manufacturing, engineering depth, systems knowledge, supply chain and network. Dede Goldschmidt, senior vice president of Samsung Electronics and head of the Samsung Semiconductor Innovation Center, said Euclyd's approach addresses real constraints in AI data centers. The company plans to use the proceeds to expand its engineering team, advance its silicon and systems development, and build out ecosystem partnerships, and Kastrup said it expects to launch its physical chip systems in 2028 and serve thousands of enterprise customers by 2030, though Euclyd has not yet demonstrated its chip systems in large-scale commercial environments. The funding arrives as investment in AI chip alternatives to Nvidia has grown, with Nvidia's share of the inference chip segment estimated between 60% and 75% in 2025 facing growing pressure from custom silicon developed by hyperscalers and startups alike.

EUCLYD has signed a Series A financing round of more than €200 million, the Eindhoven-based semiconductor systems company announced. The round is co-led by Samsung, Somerset Capital Partners, Scaleup Europe Fund, which is managed by EQT, and Innovation Industries, with participation from EIFO, the Export and Investment Fund of Denmark, imec.xpand, Brabant Development Agency (BOM), and Quadri. Peter Wennink, former President and CEO of ASML, joins EUCLYD as Chairman of the Board. The financing will expand EUCLYD's engineering organization, accelerate its silicon and systems roadmap, strengthen ecosystem partnerships, and prepare the company for commercial deployment across enterprise, sovereign, and hyperscale AI markets. At the center of EUCLYD's roadmap are craftwerk, which it describes as the world's first agentic AI silicon, and craftwerk station CWS, which it calls the world's lowest-power exascale AI factory.

A September 14 selloff treated slower frontier AI development as bad news for chips, but Reuters Breakingviews argued that a moderation in giant model-training runs could redirect part of this year's roughly $1 trillion AI investment toward inference, where existing models answer queries and run agents. That shift would create a different contest for Intel and Nvidia, because inference does not erase GPUs but broadens the bill of materials. Intel's opportunity is the server CPU, since production AI applications need orchestration, memory, networking and general-purpose compute around accelerators, and Intel reported $6.3 billion of Data Center and AI revenue in its latest quarter, up 59% year over year. Execution remains the risk: Intel is spending heavily to rebuild manufacturing leadership and still generated negative adjusted free cash flow in its latest quarter, and inference growth helps only if Intel keeps enough share and margin against AMD, Arm-based CPUs and specialized accelerators. Nvidia remains the strongest counterargument to a simple rotation thesis, since its GPUs run both training and inference and CUDA plus newer inference-focused systems give it a path to monetize deployed AI even if frontier training slows, with the downside relative rather than existential. Insider Monkey's database counted 138 hedge funds holding Intel in Q2 2026, up from 112 in Q1, with AQR Capital Management owning 10,742,567 shares after trimming its position 7%, while Nvidia rose to 285 funds from 275 and Fisher Asset Management increased its stake 3% to 90,935,947 shares, filings that predate the September 14 safety-driven market reaction. Intel had 152,248,752 shares sold short on August 31, equal to 3.02% of float, with 1.69 days to cover.

Cerebras Systems is expanding its technology roadmap and customer base to strengthen its position in the AI inference market, targeting coding, agentic AI, security and enterprise workloads as it competes with NVIDIA and Alphabet. Core cloud and other services revenues surged 287% year over year to $127.7 million in the second quarter of 2026. The company expects to double system speed annually over the next several years and increase throughput by more than 20 times over the next 18 months, while its disaggregated inference architecture with AMD combines Cerebras systems with AMD Helios racks to raise throughput by up to five times. Its collaboration with Amazon Web Services is expected to bring disaggregated inference to Amazon Bedrock in the first quarter of 2027, and Cerebras supports OpenAI's GPT-5.6 Sol at a speed that is 10 times faster, with first revenues from other hyperscalers expected around mid-2027. Cerebras signed six deals worth more than $30 million each in the second quarter and added customers such as Cognition, Lovable and Figma in AI coding, while Block, AlphaSense and GSK use its fast inference for agentic workflows and CrowdStrike represents a move into AI-powered cybersecurity. The company has more than 600 MW of data-center capacity live or contracted through the end of 2027, and manufacturing capacity is expected to rise more than tenfold in 2026, supported by secured TSMC wafer supply.

Piper Sandler initiated coverage of Intel with a neutral rating and a $110 price target, even as the firm projects the chipmaker can sustain annual revenue growth in the high teens through 2030. Intel's top line was flat last year at $52.9 billion, and consensus estimates now project a 19% increase in 2026 to $63 billion. The expected acceleration is driven by rising server CPU demand for AI data centers running inference and agentic workloads, a shift that Advanced Micro Devices estimates is moving the server CPU-to-GPU ratio toward 1:1 from 1:4 or 1:8. Intel is reportedly sitting on a server CPU order backlog of over six months, and server CPU prices have already risen 10% to 35%, with DigiTimes reporting the company plans a further 10% price increase next month. Bank of America expects the server CPU total addressable market to grow almost 5x between 2025 and 2030, reaching over $170 billion by the end of the decade.

Nvidia is letting rival AI accelerators into its racks, as Reuters reported September 10 that inference-chip startup d-Matrix will use Nvidia's NVLink Fusion technology to connect its next-generation Raptor processors directly into Nvidia data-center racks, with the systems expected to become available in 2027. Astera Labs is working with d-Matrix on custom connectivity solutions to move data rapidly across the system, creating a three-layer architecture in which d-Matrix supplies the accelerator, Nvidia supplies the rack-scale interconnect technology and architecture, and Astera helps solve the data-movement problem between components. For Nvidia, the bull case is that NVLink can become more valuable even when Nvidia does not sell every accelerator, shifting it from dominating GPUs to controlling an important system standard, while the bear case is that NVLink Fusion deliberately makes alternative accelerators easier to deploy and could cost Nvidia some GPU unit share. For Astera Labs, heterogeneous systems may be an even cleaner opportunity, since more accelerators, memory pools, CPUs and switches create more connectivity bottlenecks for its retimers, fabric switches and other connectivity products, though its valuation already assumes substantial AI infrastructure growth. Hedge-fund positioning turned more bullish on Astera Labs in the second quarter, with 73 hedge funds tracked in the stock versus 53 in the first quarter, while Nvidia rose to 285 hedge funds from 275, and short interest stood at roughly 6.6% of Astera's float on August 14 versus about 1.2% for Nvidia, though ALAB short interest fell from the previous reporting period.