The Information's Exclusive Scoop and the Full Story of the Vertex AI Leak

On the night of April 14, 2026, The Information — one of the most trusted paid media outlets in the tech industry — reported, citing sources with insider knowledge of Anthropic's internal affairs, that "Anthropic is preparing to launch its next-generation flagship model, Claude Opus 4.7, along with an AI design tool, as early as this week." The report spread instantly throughout Silicon Valley not merely as a leak story, but as major news with the potential to reshape the balance of power in the AI industry.

Two days later, on April 16, a user discovered an undisclosed model ID — "anthropic-claude-opus-4-7" — on Google Cloud Vertex AI's quota management page and published it along with a screenshot. Vertex AI is Google's enterprise AI platform and one of the infrastructure foundations through which Claude models are offered under a strategic partnership with Anthropic. The discovery attracted significant attention as technical corroboration of The Information's reporting.

In fact, an important precursor had already emerged in late March 2026. Source maps included in the Claude Code npm package were inadvertently left publicly accessible, temporarily exposing approximately 500,000 lines of source code. Boris Cherny, the head of Claude Code, acknowledged it as "a simple developer mistake," but the leak revealed internal references to Opus 4.7, Sonnet 4.8, and a next-generation model family referred to as "Mythos." Specifically, these model names appeared in a list of banned version strings for a feature called "Undercover Mode," and internal codenames were also uncovered: "Fennec" (Opus 4.6), "Capybara" (a new model tier positioned above Opus), "Numbat" (an undisclosed model), and "Tengu" (the project codename for Claude Code). Additionally, 44 feature flags were identified, revealing the existence of unreleased features such as background agents and voice mode.

On the prediction market Polymarkets, as of April 16, an implied probability of 79% was assigned to Claude 4.7 being released by that same date, with a 98% probability shown for a release by the end of June. Anthropic itself has issued no official comment on these reports or leaks, but with multiple independent sources consistently pointing in the same direction, an imminent release is considered virtually certain.

Technical Specifications and Expected Performance of Claude Opus 4.7

Evolution from the Previous Generation Opus 4.6

Claude Opus 4.6 was released on February 5, 2026, achieving industry-leading performance in both coding and reasoning. It delivered scores of 80.8% on SWE-bench Verified, 65.4% on Terminal-Bench 2.0, and 72.7% on OSWorld, and its groundbreaking pricing — a 1-million-token context window at standard rates ($5 input / $25 output per million tokens) — also drew significant attention. Opus 4.7 is an incremental upgrade built on 4.6, but multiple analytics sites predict "a 15–25% performance improvement on coding benchmarks." There are indications it could reach scores of 85–90% on SWE-bench Verified, which, if realized, would reclaim the top position on that benchmark.

Specific technical improvements cited for Opus 4.7 include enhanced accuracy in multi-step reasoning, stronger autonomous long-horizon task execution, and native multi-agent coordination support. Anthropic has been actively developing multi-agent collaboration capabilities in recent years through its Agent SDK and Claude Code, and Opus 4.7 may optimize these at the model level. The "Coordinator Mode" (multi-agent swarm orchestration) discovered in an npm leak is precisely in line with this direction.

Some analytics sites have also mentioned the possibility that Opus 4.7 could adopt a Mixture-of-Experts (MoE) architecture, though this remains unconfirmed and is assessed at roughly 60–70% confidence.

The Outlook for a 2-Million-Token Context Window

The current Claude Opus 4.6 and Sonnet 4.6 offer a 1-million-token context window at standard pricing. This became generally available (GA) in March 2026, and the elimination of the traditional surcharge for long-context usage has led to explosive growth in use cases such as large-scale codebase analysis and lengthy document processing.

There are observations suggesting that Opus 4.7 plans to expand to a 2-million-token context window for higher tiers such as Max and Enterprise plans. Given that Google's competing model, Gemini 2.5 Pro, can already natively handle 1 million tokens and has hinted at further expansion, it is highly rational from a competitive strategy standpoint for Anthropic to establish "2 million tokens" as the new threshold. Expanding the context window is not merely a numerical increase — it opens the door to new use cases such as bulk analysis of large enterprise repositories, review of legal documents spanning hundreds of pages, and agentic operations that retain multiple days' worth of conversation history.

From an API pricing perspective, one of the greatest concerns in the developer community is whether Opus 4.7 will maintain the policy established with Opus 4.6 of "standard pricing regardless of context length." The prediction site Apiyi.com forecasts Opus 4.7 API pricing at $15–20 for input and $75–100 for output (per million tokens), though the possibility that Anthropic will maintain its current $5/$25 price range to stay competitive has not been ruled out.

Native Video Input and 60fps Support

At the frontier of multimodal AI, the ability to process video input has become a key differentiator among companies. Current Claude models support text and image input, but native video input is not yet supported, and developers have resorted to indirect methods such as frame extraction via FFmpeg or third-party toolkits. On GitHub, a feature request (Issue #32130) calling for the addition of native video analysis capabilities to Claude Code has sparked active discussion.

Coinciding with the release of Opus 4.7, the introduction of native video input has been observed. The specific 60fps support is strongly aimed at enterprise use cases that demand high frame rates, such as industrial video analysis, sports analytics, and security surveillance. Anthropic released the Claude Video API in 2026, achieving processing latency of under 500ms per frame for analysis tasks, with support for H.264 and VP9 codecs and resolutions up to 4K. The API adopts a usage-based pricing model of $0.06 per minute for analysis and $0.18 for generative video features, and takes the strategic positioning of providing "explainable and controllable video analysis" for enterprise customers.

Anthropic CTO Jared Kaplan has said of Claude's video capabilities: "Real-time performance, robust privacy features, and transparency in model outputs — these are indispensable elements for enterprise adoption." If native video input with 60fps support is realized, it will establish a clear advantage over competitors OpenAI's GPT-4o and Google's Gemini 2.5 Pro, particularly in the areas of enterprise security and real-time video analysis.

LumiAI, an accessibility startup based in London, uses the Claude Video API to provide real-time sign language interpretation at live events and reports that it has "reduced latency by 40% compared to conventional pipelines." In the security space, enterprise integrator VigilantEye leverages Claude Video analysis for anomaly detection in retail environments, detecting potential threats such as theft and security incidents in near real time.

The Full Picture of AI Design Tool "Project Prism / Claude Design Studio"

An Approach That Overturns Conventional Design Tools From the Ground Up

According to reporting by The Information, the AI design tool Anthropic is preparing will allow users to generate websites, landing pages, presentations, and product mockups all at once simply by entering a natural language prompt. This approach is fundamentally different from how Adobe Firefly or Figma AI function as "assistants to trained designers." Anthropic's tool replaces "the very starting point of design itself," enabling users without design expertise to produce professional-quality outputs.

This concept is evident from a screenshot of "Claude Builder" that leaked on April 14, 2026. The interface, accompanied by the tagline "Let's ship something great," features template-based design, real-time app previews, integrated database and authentication tools, user analytics, and a one-click deployment button. It is said to be capable of building full-stack applications from simple prompts — from landing pages and AI chatbots to photo albums and even games like Space Invaders.

Analysis by FindSkill.ai cites "Claude Studio" as the rumored name for this design tool, predicting it will include export capabilities to Figma, HTML, React, and native formats. Given that Anthropic has already enabled the conversion of AI-generated code into editable design files via Figma integration, as well as Claude integration with Microsoft Word and PowerPoint, the design tool is likely to be deployed in a way that connects seamlessly with Anthropic's existing ecosystem.

Competitive Landscape: Positioning Against Google Stitch, Lovable, Bolt.new, v0, and Gamma

The market for AI design tools and AI app builders has reached $4.7 billion in revenue as of 2026, and is projected to expand to $12.3 billion by 2027. Several prominent players already exist in this rapidly growing market.

Stockholm-based Lovable raised a $330 million Series B in December 2024 with backing from NVIDIA, Salesforce, Databricks, and Atlassian, reaching a valuation of $6.6 billion. The company has surpassed $200 million in ARR (annual recurring revenue) and has recorded the fastest growth of any European startup in history. Bolt.new, operated by StackBlitz, has achieved $40 million in ARR, demonstrating the viability of the new paradigm of browser-based AI development. Vercel's v0 specializes in frontend component generation and, while it does not offer backend functionality, has secured a unique position through its integration with the Next.js ecosystem. Google's Stitch is available for free in its Labs stage and, as a design system leveraging Gemini 2.5 Pro, has received high praise in the developer community — particularly in combined workflows with Claude Code. In the presentation space, Gamma leads in generating slide decks from natural language, while the segment is already crowded with Beautiful.ai, Tome, Canva's AI features, and Microsoft Copilot in PowerPoint, among others.

What makes Anthropic's design tool decisively different is its structure: a foundation model company building first-party productivity tools on top of its own AI foundation. As PYMNTS analysis points out, Anthropic is pursuing a "transition from language model provider to full-stack AI studio," aiming for a world where Claude builds and deploys entire products. This is a fundamentally different approach from conventional workflows that use separate SaaS products at each layer of the toolchain, and this structural threat is precisely what investors are most wary of.

Stock Market Impact on Design Tool-Related Equities

On April 14, when The Information's report was published, design tool and web-building equities fell across the board. FIG, the stock of the recently IPO'd Figma, dropped 6% and suffered the largest blow. Wix fell 4.7%, Adobe fell 2.7%, and GoDaddy fell 3%. BTIG initiated coverage of Adobe and Figma at this timing, assigning both stocks a "Neutral" rating and declining to set price targets. A comment from a BTIG analyst — "At the core is concern that AI will fundamentally transform how content is created, and we are not ready to fully commit" — succinctly captures the mood of the market.

Sentiment analysis on Stocktwits shows that retail investor sentiment for ADBE, GDDY, and FIG remains in "bullish" territory, with a certain number of investors viewing the decline as a short-term buying opportunity. One user predicted "Claude will put Adobe and Figma out of business," while another downplayed the impact by stating "Figma's product is better than anything on the market." More significant, however, are the macro-level figures. The S&P 500 Software & Services index has fallen nearly 26% in 2026, with structural concerns that AI tools are eroding demand for conventional software products weighing on the sector as a whole.

Worth noting, however, is the possibility that Anthropic's design tool may not entirely negate its relationships with existing partners. Of the six launch partners for the Claude Marketplace that Anthropic launched in March 2026, two were Replit and Lovable. There is also a view that Anthropic is positioning itself not as a "replacement" for vibe-coding tools but as their "foundational platform," and the balance between platform strategy and product strategy will be the key that determines how events unfold going forward.

Silicon Valley VCs' Reaction and the Shock of an $800 Billion Valuation

Soaring Valuation and the VC Investment Offensive

According to reports by Bloomberg and Business Insider, Anthropic has been flooded with investment offers from multiple VCs for a new funding round at a valuation exceeding $800 billion. This means the valuation has more than doubled in just two months from the Series G in February 2026 (which raised $30 billion at a pre-money valuation of $350 billion). According to TechCrunch, Anthropic has "for now" held off on accepting these offers.

This valuation would surpass — or at least rival — that of competitor OpenAI (whose latest valuation is said to be in the $300 billion range). Behind the Silicon Valley VC community's rush to invest in Anthropic lies the company's remarkable revenue growth. Anthropic's annualized revenue run rate surged from $9 billion at the end of 2025 to $30 billion as of the end of March 2026, recording approximately 1,400% year-over-year growth. This pace of growth is what InvestorPlace describes as "an unprecedented 10,000% revenue growth rate in tech history that could make the IPO the largest of 2026."

Key Investors and Strategic Partner Developments

Anthropic's investor composition is breaking with conventional practices in the Silicon Valley VC industry. In January 2026, Sequoia Capital — also a major investor in OpenAI — was reported by the Financial Times to have made a large investment as part of Anthropic's Series G. Sequoia also has stakes in OpenAI and Elon Musk's xAI, openly flouting the VC industry taboo of "avoiding investments in competing companies." This move is interpreted as a manifestation of a new investment strategy: rather than betting on a single winner in the AI space, spreading a portfolio across multiple foundation model companies.

The Series G is led by Singapore's GIC and U.S.-based Coatue, each contributing $1.5 billion, with Sequoia Capital reportedly planning an investment described as "in the billions." On the strategic investor side, Amazon and Alphabet/Google hold the largest stakes, and as competition over AI infrastructure intensifies, both companies have deep strategic interests in Anthropic's success.

In the 2026 AI-focused VC rankings, Thrive Capital holds the overall top position, with Lightspeed and Andreessen Horowitz (a16z) close behind. a16z is known for its strong technical due diligence and a broad yet selective AI portfolio, while Benchmark Capital takes a highly selective approach focused on deep technical partnerships at the infrastructure layer. The fact that these top-tier VCs are collectively expanding their investments in AI foundation model companies has become one of the defining themes of VC investment trends in 2026.

The Path to IPO

Anthropic has its sights set on an IPO in the second half of 2026, and according to Bloomberg, Goldman Sachs, JPMorgan, and Morgan Stanley have entered early-stage discussions as candidates for lead underwriter. The listing is anticipated as early as October 2026, with the potential to raise more than $60 billion — which would make it one of the largest technology IPOs in history. Wilson Sonsini, the law firm behind the IPOs of Google and LinkedIn, is handling structural and regulatory preparations, with a practical timeline envisioning an S-1 filing in late summer 2026, followed by a two-to-three-week roadshow and an October listing.

It is worth noting that OpenAI is also targeting a listing in the second half of 2026, with Bloomberg reporting that "the two companies are racing each other to go public first in order to capture institutional investors' allocation budgets." With SpaceX's roadshow potentially beginning as early as June 2026, if the OpenAI and Anthropic IPOs are also concentrated in the second half of the year, there are warnings that the three companies could collectively exhaust demand for AI-related IPOs in 2026.

Claude Code's Explosive Growth and Enterprise Platform Strategy

The Battle for Developer Tools Supremacy

Since its GA in May 2025, Claude Code has expanded its market share at a remarkable pace, overtaking GitHub Copilot in just 8 months to become the top AI coding assistant. It has grown to accumulate over 81,000 GitHub Stars and generate more than 135,000 GitHub commits per day (equivalent to 4% of all public GitHub commits worldwide), with some forecasts projecting that figure to reach 20% by year's end. 46% of developers name Claude Code their "most loved tool," far ahead of second-place Cursor (19%) and GitHub Copilot (9%).

On the revenue side, Claude Code's ARR has reached $2.5 billion, doubling since early 2026. Business subscriptions have quadrupled since January, with enterprise users now accounting for more than half of Claude Code revenue. The speed at which it reached a $1 billion run rate just six months after GA is unprecedented in SaaS history.

Claude Managed Agents and Marketplace

Anthropic officially launched "Claude Managed Agents" on April 8, 2026 — a cloud service that provides AI agent sandboxing, orchestration, and governance for enterprise customers. The company's claim of "shortening development workflows from months to weeks" remains to be validated by early adopters such as Notion, Rakuten Group, and Asana. Pricing is charged at $0.08 per agent runtime hour, in addition to Claude model usage fees.

VentureBeat's analysis describes Claude Managed Agents as "a new one-stop shop that carries the risk of vendor lock-in," while also pointing to Anthropic's strategic intent to capture value across multiple layers and build an environment that generates network effects. Anthropic has secured a majority share among first-time AI buyers, and as OpenAI's share declines, the balance of power in the enterprise market is clearly shifting.

Claude Mythos and "Project Glasswing" — The Most Powerful Model That Won't Be Released

If Opus 4.7 is Anthropic's "commercial flagship," Claude Mythos occupies the position of "frontier research output." After its existence was revealed by an npm leak in late March 2026, Anthropic officially announced Mythos Preview on April 7th, but made the unusual decision not to release it to the general public.

Mythos Preview achieved an SWE-bench score of 93.9%, a leap of 13.1 percentage points from Opus 4.6's 80.8%. On a benchmark where improvements of 2–3 points were previously considered "meaningful progress," this is literally a "step change." Internal testing confirmed its ability to autonomously discover and exploit zero-day vulnerabilities across all major operating systems and web browsers—reportedly detecting a 27-year-old bug in OpenBSD and a 16-year-old vulnerability in FFmpeg's H.264 codec.

Anthropic provides access to Mythos exclusively through a consortium called "Project Glasswing." Participating organizations include an impressive roster—AWS, Apple, Cisco, CrowdStrike, Google, JPMorgan Chase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks—with Anthropic committing $100 million (approximately ¥15.9 billion) in usage credits. Mythos's system card, exceeding 240 pages, is the most extensive safety document Anthropic has ever published, and is governed under the RSP 3.0 (Responsible Scaling Policy 3.0) framework adopted by the company in February 2026.

The Council on Foreign Relations has published an analysis piece titled "Six Reasons Claude Mythos Marks a Turning Point for AI and International Security," warning that shifts in the offense-defense balance in cybersecurity carry geopolitical implications. The asymmetric release strategy—Opus 4.7 as "the most powerful model available to the public" versus Mythos as "the most powerful model held in containment"—presents a new paradigm in the discourse around AI safety.

Analysis of Coverage and Tone in Major Media

Tech Specialty Media

The Information (April 14) broke the exclusive scoop, triggering a wave of coverage across tech-focused outlets. Decrypt referenced the existence of Mythos under the headline "Anthropic Preps Opus 4.7 and a Full-Stack AI Studio—But It Has Something Scarier Too," highlighting the two-tier structure in which Anthropic would deploy Opus 4.7 as its commercial layer while offering Mythos exclusively to national security partners. TechCrunch reported with the nuance that Anthropic is "for now" turning down offers at an $800 billion valuation, suggesting the company may be maneuvering to push its valuation even higher ahead of an IPO.

Dataconomy reported "Anthropic to Launch Claude Opus 4.7 This Week" and cited claims that Opus 4.6's performance had been deliberately throttled—speculation that this was a staged "generational handoff" timed to the new model's release. VentureBeat analyzed Anthropic's overall platform strategy alongside the launch of Claude Managed Agents, offering a cautious assessment: "a new one-stop shop for the enterprise, but with vendor lock-in risks."

Financial & Investment Media

Bloomberg reported "Anthropic Attracts Investor Offers at $800 Billion Valuation," emphasizing the company's reserved stance toward those offers. PYMNTS.com published a detailed analysis titled "Anthropic's New Design Tools Rival Adobe and Figma," noting that Figma holds an estimated 80–90% share of the UI/UX design market while arguing that Anthropic's approach could fundamentally disrupt that market structure through the democratization of design. InvestorPlace focused on the possibility that Anthropic's "10,000% revenue growth" could produce 2026's biggest IPO, while SaaStr analyzed the strategic implications of Anthropic surpassing OpenAI in revenue.

GuruFocus reported in detail on the impact to shares of Adobe, Figma, Wix, and GoDaddy, and MarketScreener debated whether Anthropic's model-plus-tools strategy could become a "software industry killer." CryptoIntegrat reported "Anthropic Claude Opus 4.7 and AI Design Tools Imminent, Adobe and Figma Stocks Drop 2%," illustrating that even crypto media was paying close attention to the ripple effects of this news.

Competitive Landscape of the AI Market and Strategic Positioning of Each Company

Market Share and User Trends

As of April 2026, ChatGPT holds a 68% share of the AI chatbot market, a significant drop from 87.2% a year earlier. Google Gemini has surged from 5.4% to 18.2%, and together the two form a "duopoly" controlling 86.2% of the market. Claude's overall share remains at 2–4.5%, yet in enterprise head-to-head comparisons it is said to win roughly 70% of deals against OpenAI, reflecting a clear divergence in strategy between volume and quality.

From a user engagement perspective, Claude users average 34.7 minutes per daily session, the longest of any major AI platform. ChatGPT commands an overwhelming consumer base of 800–900 million weekly active users and 67.7 million monthly active users (U.S. only), while Claude has carved out a well-defined niche among "developers and writers who demand precision and coding capability."

Model Performance Comparison

In coding benchmarks, Grok 4 leads with a raw SWE-bench score of 75%, followed closely by OpenAI's GPT-5.4 at 74.9% and Claude Opus 4.6 at over 74%, making it a tight three-way race. However, if Claude Opus 4.7 achieves the projected score of 85–90%, these rankings will shift dramatically. Mythos's 93.9% stands in a class of its own, but is excluded from direct comparisons as it is not publicly available.

Google is leveraging its distribution power (Android, Google Search, Chrome) to accelerate Gemini adoption, with Gemini 2.5 Pro differentiating itself through a one-million-token context window and native multimodal reasoning. OpenAI maintains its dominant position through consumer scale (ChatGPT Plus, Enterprise) and sustains its technical competitiveness through ongoing improvements to the GPT-5 series. Anthropic is pursuing a more targeted strategy centered on enterprise accuracy and developer tooling (Claude Code and integrations with Cursor and Windsurf).

Future Outlook: When and What Will Happen

Short-term (April–May 2026)

The most imminent event is the official release of Claude Opus 4.7 and the design tools. Since the "within this week" (April 14–18) window reported by The Information is closing soon, an announcement on April 17 or early the following week is considered most likely. Polymarket odds show near-certain probability of a release within April. Once the release is confirmed, API compatibility is expected to be highly compatible with Opus 4.6, meaning migration costs for existing users should be minimal. For the design tools, a phased rollout is anticipated — early access for Team and Enterprise plans first, followed by expansion to the Pro plan.

In May, an official announcement of Project Glasswing is expected in San Francisco, and the full deployment of Mythos Preview to security partners is set to begin. Early data on Claude Managed Agents adoption will also start to emerge, initiating verification of the company's claim that it "speeds up development workflows by 10x."

Mid-term (June–October 2026)

Moving into summer, the release of Sonnet 4.8 comes into view. This model, whose existence was confirmed via an npm leak, is expected to follow a cost-performance optimization path distinct from Opus 4.7. Additionally, feature expansion and API availability for the design tools will advance, and the formation of a third-party ecosystem will begin in earnest.

The biggest event is the IPO anticipated in October. If the S-1 filing takes place in late summer, the market will gain access to Anthropic's detailed financial information for the first time. If figures such as $30 billion in annualized run-rate revenue, over 1,000 enterprise customers paying more than $1 million annually, and surpassing OpenAI in the enterprise market are officially confirmed, IPO demand is expected to be extremely strong.

Long-term (Late 2026–2027)

CEO Dario Amodei has predicted that AI capabilities on the level of a "nation of geniuses within a data center" will be realized within one to three years (around 2026–2027), and the existence of Mythos suggests this prediction is gaining credibility. Anthropic's infrastructure strategy — compute diversification across three pillars of Google TPUs, Amazon Trainium, and NVIDIA GPUs, along with a 3.5-gigawatt compute agreement with Google and "Project Rainier" with Amazon — is positioned as the foundational buildout needed to support this ambitious vision.

In the AI design tools market, the key watch point is how much of Figma's 80–90% UI/UX design market share will be eroded. Whether Adobe's per-application billing model, which underpins its annual revenue of $23.77 billion (approximately ¥3.7794 trillion), remains sustainable against lower-cost AI-native tools will first become visible as concrete data during the earnings season in the second half of 2026.

Investment Implications from a VC Perspective

For the Silicon Valley VC community, the release of Opus 4.7 and the design tools affects investment decisions on multiple layers.

First, the appeal of direct investment in Anthropic itself has grown even stronger. The $800 billion valuation may seem astronomical at first glance, but a multiple of 26.7x against a run-rate revenue of $30 billion is by no means excessive for an AI company. Even compared to the P/S ratio of OpenAI's most recent valuation (in the $300 billion range) against estimated revenue ($25 billion), this can be justified when factoring in Anthropic's growth rate.

Second, Anthropic's platform expansion demands an impact analysis on portfolio companies. With the introduction of design tools, startups such as Lovable, Bolt.new, v0, Gamma, Beautiful.ai, and Tome are being forced to redefine their own value propositions. However, as the Claude Marketplace strategy suggests, if Anthropic chooses to function as a platform layer, these startups may also find a new growth path through differentiation built on top of Claude. The comment by Elena Verna, Head of Growth at Lovable — that "the real competition is not other vibe-coding startups, but Big Tech platforms" — aptly captures this structural shift.

Third, Anthropic's move toward a full stack blurs the line between "investing in AI infrastructure" and "investing in AI applications." Traditionally, VCs have made investment decisions by clearly distinguishing the infrastructure layer (model companies) from the application layer (SaaS startups), but in a world where Anthropic deploys a full stack comprising design tools, Managed Agents, Claude Code, and Marketplace, this dichotomy no longer functions. In AI VC investment in 2026, the reassessment of "platform risk" is becoming the single most important theme.

Investment Implications from a VC Perspective

For the Silicon Valley VC community, the release of Opus 4.7 and the design tools affects investment decisions on multiple layers.

First, the appeal of direct investment in Anthropic itself has grown even stronger. The $800 billion valuation may appear astronomical at first glance, but a 26.7x multiple against a run-rate revenue of $30 billion is by no means outrageous for an AI company. Even compared to the P/S ratio of OpenAI's most recent valuation (in the $300 billion range) against estimated revenue ($25 billion), this is a level that can be justified when factoring in Anthropic's growth rate.

Second, Anthropic's platform expansion demands an impact analysis on portfolio companies. With the introduction of design tools, startups such as Lovable, Bolt.new, v0, Gamma, Beautiful.ai, and Tome are being forced to redefine their own value propositions. However, as the Claude Marketplace strategy suggests, if Anthropic chooses to function as the platform layer, this also opens a new growth path for these startups — differentiation built on top of the Claude foundation. Elena Verna, Head of Growth at Lovable, captured this structural shift precisely when she said, "The real competition isn't other vibe-coding startups — it's Big Tech platforms."

Third, Anthropic's move toward a full-stack model blurs the boundary between "investing in AI infrastructure" and "investing in AI applications." Traditionally, VCs have made investment decisions with a clear distinction between the infrastructure layer (model companies) and the application layer (SaaS startups), but in a world where Anthropic deploys a full stack comprising design tools, Managed Agents, Claude Code, and Marketplace, this binary no longer functions. In AI VC investment in 2026, the reassessment of "platform risk" is becoming the most critical theme.

Sources