Skip to content
Christmas & Holiday Lights 2026Explore the community map

Agentic Silicon & Frontier Platforms

Nvidia Unveils Vera CPU as Claude Adds Plaid and OpenAI Eyes 2027 IPO

Nvidia debuts the 176-core Vera CPU delivering 1.6x more agents per watt in a 100-megawatt data center, Anthropic rolls out Plaid-connected Personal Finance for Claude Pro, Max, and Ultra users, OpenAI President Greg Brockman signals 2027 IPO readiness after 70% Q3 growth, and Meta launches Muse Voice across the U.S.

By San Ramon Times · Syndicated from Lazy Tensor
Syndicated reporting from Lazy Tensor · Lazy Tensor
Published · 4 min read

Editorial crosshatch print illustration on cream paper showing a silicon wafer processor connected to a balance scale and soundwave rings
An editorial print illustration depicting a geometric multi-core processor wafer linked to an architectural balance scale and acoustic wave rings. · Lazy Tensor / Editorial Visual Desk
Your listening Queue & catch up
Unheard stories first, timed at 1×. Your selections stay first.

Add a story or build a catch-up queue to start listening.

Hardware architects and frontier artificial intelligence labs rolled out dedicated server silicon for autonomous software agents, consumer financial analysis integrations, real-time voice avatars, and updated public-market timelines on Thursday.

Nvidia Debuts 176-Core Vera CPU Purpose-Built for Autonomous AI Agents

Nvidia introduced the NVIDIA Vera CPU for agentic workloads, detailing how the processor handles the branching logic, sandbox orchestration, database queries, and tool-calling loops that accompany multi-step AI agents. Across a 100-megawatt data center footprint, Nvidia reported that the Vera CPU delivers 1.6x more agents per watt than the fastest x86 processor and 2x faster agent speed than the most energy-efficient x86 CPU.

Early infrastructure partners reported substantial throughput gains across production benchmarks:

  • Perplexity: Achieved 1.5x faster agentic coding workloads and higher accuracy at scale.
  • DeepInfra: Recorded up to 2.2x speed improvements on agentic applications.
  • ClickHouse: Scaled its real-time analytics database linearly across all 176 cores of the Vera CPU.
  • DuckLabs: Measured a 1.5x performance boost on TPC-H SF1000 analytical benchmarks compared with Intel Xeon 6 processors.
  • Daytona: Spun up 1,000 isolated code sandboxes in 10.8 seconds on Vera compared with 88.4 seconds on x86 servers, while sustaining 12.7 jobs per second across 2,000 concurrent sandboxes.
  • Modal: Delivered 2x higher sandbox throughput per core alongside tighter tail latency under heavy multi-tenant load.

Anthropic Launches Plaid-Powered Personal Finance Tracking in Claude

Expanding into consumer financial management five months after OpenAI introduced similar capabilities in ChatGPT in May, Anthropic launched a Personal Finance integration in beta for U.S. subscribers on Claude Pro, Max, and Ultra plans across web, desktop, and mobile.

Through a read-only connection powered by financial data network Plaid, users can link bank accounts, credit cards, loans, and investment portfolios directly to Claude. Once connected, Claude can inspect up to 2 years of transaction history, account balances, the last 4 digits of account numbers, credit limits, and annual percentage rate (APR) interest charges to build custom budgets, flag recurring subscriptions, and model debt-payoff schedules. Anthropic emphasized that the Plaid integration is strictly read-only—preventing the assistant from moving money or executing trades—and that personal financial records are never used to train foundation models.

OpenAI President Greg Brockman Signals Potential 2027 IPO Readiness

Speaking at a Wall Street Journal tech conference on Thursday, OpenAI President Greg Brockman said the company could be ready for an initial public offering in 2027, contingent on market conditions and internal operational readiness.

Brockman disclosed that OpenAI posted 70% quarterly annualized revenue growth during the third quarter of 2026. Enterprise subscriptions and API deployments have doubled as a share of the business, rising from roughly 30% of revenue to represent the majority of OpenAI's total revenue. Brockman pointed to OpenAI's enterprise deployment partnership with JPMorgan Chase and rapid corporate adoption of its autonomous software engineering agent Codex as primary drivers of the shift toward enterprise recurring revenue.

Meta Rolls Out Muse Voice Video Chat Avatars to U.S. Users

On the consumer agent front, Meta CEO Mark Zuckerberg and Chief AI Officer Alexandr Wang announced that Muse Voice is now available to all users in the United States. First previewed at Meta's Connect developer conference in September, the feature uses Meta's new Muse Realtime Avatar and Muse Realtime Voice models to let users place live voice and video calls to their personal Muse AI agent.

By default, Muse appears as an animated character nicknamed Jolly that responds with real-time facial expressions and gestures matched to conversational tone, though users can customize the avatar to resemble a person, animal, or fictional character—as Wang demonstrated on X with a video of his fluffy white dog avatar advising him on social media memes. Muse Voice is live today across mobile (including the iOS app), web, and Mac, and Meta plans to extend support to Meta AI Glasses and its upcoming Tamagotchi-style "Muse Charm" wearable hardware device before year-end.

Sources

Comments (0)

No comments yet. Be the first to share a thought on this story.