• Skip to main content
  • Skip to secondary menu
  • Skip to footer

Technologies.org

Technology Trends: Follow the Money

  • Technology Events 2026-2027
  • Sponsored Post
  • Technology Markets
  • About
    • GDPR
  • Contact

Meshy Raised $400 Million at a $1.5 Billion Valuation and Announced It Two Different Ways

July 25, 2026 By admin

Meshy, which generates usable 3D models from a line of text or a single photograph, closed a Series B of nearly $400 million this week. The company calls it the largest round ever raised by a business built specifically for AI 3D generation, and by the available evidence that is true. It is also the company’s first publicly disclosed valuation since it was founded in 2023.

The more interesting fact is that the announcement came in two versions, and they do not say the same thing.

The English release omits what the Chinese release leads with

The English-language wire announcement carries a Silicon Valley dateline and credits the round to a group of leading global investors. It names none of them. At least one funding tracker recorded the backers as undisclosed on that basis.

The Chinese-language coverage, published a day earlier, names all seven. IDG Capital, Matrix Partners China and Monolith co-led. Granite Asia, HongShan — the firm that was Sequoia China until the 2023 split — BAI Capital and Source Code Capital all oversubscribed their pro-rata allocations. Chinese trade press describes Meshy as Beijing-based and identifies the founder as Hu Yuanming, a graduate of Tsinghua’s Yao Class who went on to an MIT doctorate in computer graphics. In English he goes by Ethan Hu, and Meshy’s own release places the company’s headquarters in Silicon Valley. His LinkedIn says Sunnyvale.

None of these are contradictions in the strict sense. Companies have engineering in one country and incorporation in another all the time, and a founder can hold both a Tsinghua degree and a California address. But the pattern of disclosure is worth naming plainly: the investor roster that constitutes the single most notable thing about this round is present in one language and absent in the other. Whatever the intent, the effect is that an English-reading enterprise buyer doing a five-minute check on a new vendor will not encounter it.

The valuation is partly a rounding decision

The Chinese announcement puts the post-money valuation above 10 billion yuan. Converted at the rate one outlet used, that is roughly $1.38 billion. The English release says $1.5 billion.

The gap is not fraud and probably not even spin in any deliberate sense — currency conversion on a threshold figure produces exactly this kind of drift, and “over 10 billion yuan” is a genuinely round number in its home market. But it does mean the headline valuation being repeated across Western trade coverage is the more generous of two defensible readings, and that the unicorn framing rests on which side of the exchange rate you stand on.

The dilution question is more substantive and remains open. If $1.5 billion is post-money, a $400 million round represents something close to 27 percent of the company sold in a single Series B. That is heavy for the stage, particularly for a company its backers are describing as a category winner. It is the sort of structure that shows up either when a founder is buying a very long runway on purpose, or when the syndicate had leverage. The disclosures do not say which.

Twelve times growth from an unstated base

Meshy reports more than 12 million registered users, over 100 million models generated, and annual recurring revenue growing roughly twelvefold year over year. It does not report the revenue.

A multiple without a base is a marketing figure. Twelvefold growth on $2 million is a different company from twelvefold growth on $30 million, and the omission is conspicuous in an announcement otherwise dense with specifics. The unit economics compound the ambiguity: the company’s own pitch is that a model costs about a dollar and takes about a minute. If 100 million cumulative generations is the headline usage number, the arithmetic ceiling on consumption revenue to date is not large relative to a valuation north of a billion dollars. The value has to be somewhere else — in enterprise seats, in volume licensing, in the API — and the company has not said where.

Registered users are the softest number in the set. Free-tier signups for a tool that produces a novelty output on the first click accumulate quickly and convert poorly. Twelve million of them tells you the top of the funnel works. It tells you nothing about the rest of it.

The strongest claim is the least glamorous one

Set the financing aside and Meshy has a real technical argument, and it is narrower than the category it sits in.

Luma AI and World Labs are chasing world models — photorealistic reconstruction, coherent generated environments, the spatial-intelligence thesis. Meshy is optimizing for watertight, manufacturable geometry and cites a 97 percent slicer success rate for 3D printing. That is not a demo metric. Slicer failure is the specific reason generated meshes have historically been useless downstream: non-manifold edges, inverted normals, geometry that looks correct in a viewport and collapses the moment a printer or a physics engine touches it. Solving it is unglamorous and it is exactly the constraint that separates a toy from a production tool.

The customer list supports the narrow reading. The named partners are game studios — Nexon, NetEase Games, 37 Interactive Entertainment — and consumer 3D printer manufacturers: Bambu Lab, Creality, Elegoo, FlashForge, xTool. Those are pipeline customers with a concrete bottleneck, not enterprises buying a vision. The company also says teams inside five of the world’s ten largest technology companies by market capitalization are building with it, which is a real signal if the seats are paid and a weak one if they are engineers with corporate cards.

An entire category funded from one place

The Meshy syndicate is not an outlier. Tripo AI, its closest direct competitor, raised nearly $200 million across a Series A and A+ at the start of June and another $150 million in a Series A3 a month later, backed by Geely Capital, the gaming firms 4399 Network, Tanwan and Giant Network, and Fosun Capital. Between the two companies, the great majority of capital deployed into AI 3D generation in 2026 traces to China-headquartered or China-focused funds.

This is the part that will matter in eighteen months, and it will matter for reasons that have nothing to do with the technology. Generated 3D assets flow into game engines, into simulation environments, into manufacturing files. Bambu Lab, one of Meshy’s named printing partners, has already spent two years as a live object in American policy arguments about connected hardware and firmware control. A vendor whose asset-generation layer sits upstream of that pipeline, whose cap table is entirely Chinese, and whose headquarters is described one way in Beijing and another way in Sunnyvale is going to attract procurement questions in Western enterprises regardless of how good the meshes are.

The category’s growth projections assume the current funding pace holds. Market estimates put generative 3D at a little over $3 billion this year and near $9 billion by 2030, a figure that comes from a vendor-adjacent industry report and should be read accordingly. Whether the money keeps arriving on these terms is a separate question from whether the technology works — and on present evidence, the technology mostly works.

What the $1.5 billion is buying is less clear.

Filed Under: News

Footer

Recent Posts

  • AI in Medicine Moves From Demos to Trials With Viz.ai Heart Data, Cancer Grand Challenges AURORA Teams and a Feinstein Delirium Grant
  • Nooks Gets 500 Million Dollars From General Catalyst for AI Sales Agents as Phaze and Shackleton AI Raise Seed Rounds
  • Kalepa Raises $75 Million for Insurance AI as Capgemini and Rootstock Surveys Show Agents Still Need a Human Sign-Off
  • Visa Benchmark Finds AI Shopping Agents Fail 42.8% of Runs as FIS Survey Shows Consumers Want AI Guarding Money First
  • Braid Raises $38 Million to Put Branded eSIM Plans Inside Travel and Banking Apps
  • Cognizant Puts 25,000 Staff Behind OpenAI as IT Services Firms Become the Delivery Channel for Enterprise AI
  • Rapidus Taps HCLTech Engineers in India as Japan’s 2nm Foundry Still Hunts for Its First Signed Customer
  • XGIMI AURA 3 Projector Raises $15 Million on Kickstarter in 23 Hours as Brands Turn Crowdfunding Into Pre-Orders
  • Group14 Raises 150 Million Euros and Picks Spain for Its Next Silicon Anode Plant While Moses Lake Waits for 2027
  • Collective, Wabi and AvalonX Raise $165 Million as OPay Pulls Its $304 Million NYSE IPO on Debut Day

Media Partners

  • Market Analysis
  • BareProxy.com
  • App Coding
SovAI Joins NVIDIA Inception With Two Grid-Connected AI Factory Sites Built to the DSX Reference Design
Dropbox (DBX) Pays 350 Million Dollars for Tigris Data to Sell AI Storage as Its Core Business Stays Flat at 2.56 Billion Dollars ARR
Why Tech Stocks Are in Free Fall, AI Spending Fears and the November 5 Selloff
Semicap Pulse Reads 48.5 Steady as Chip Equipment Stocks Slump While Orders and Exports Run Hot
Memory Cycle Gauge Reads 82 Shortage as Micron Stock Breaks Trend While Supply Stays Sold Out
AI Market Analyzer Falls to 31.8 Cooling as Price Gauges Hit AI Winter and Leverage Stays Greedy
AI Capex Tracker November 2026, Hyperscalers Spend $237 Billion in One Quarter as Free Cash Flow Turns Negative
The Collingridge Dilemma
The Sum of All Fears Hits Markets as Gold Sinks to One-Year Low and Silver Slips Below $50
IBM Will Never Be the Company It Used to Be
BareProxy Adds AI Crawler Control That Tells Real GPTBot and ClaudeBot Requests From Fakes
BareProxy Adds Bot Challenges That Make Scrapers Posing as Browsers Solve a Puzzle First
BareProxy Adds Cookie Consent That Holds Back Tracking Cookies Until a Visitor Says Yes
BareProxy Adds Markdown Serving That Turns Markdown Files Into Web Pages With No Build Step
BareProxy Ships Its First Nine Plugins, From Maintenance Pages and CORS to Geo Rules and Rate Limits
BareProxy 0.2.0 Runs WebAssembly Plugins Through the Proxy-Wasm Interface
BareProxy 0.3.0 Writes Its Plugins in Rust, Where a Plugin Adds 15 Microseconds Instead of 42
BareProxy Is an Operating System for Web Traffic: A Small Core and a Long Stream of Plugins
BareProxy Plugins Go WebAssembly, Starting With RenderCache and AI Crawler Control
BareProxy Runs on Linux, macOS and Windows, and One Plugin File Runs on All of Them
Claude Code on GitLab vs GitHub: Thinner Integration and a Fifth of the Free CI Minutes
How k4i.com Publishes Posts Unattended With Small AI Models, Groq and GitHub Actions
How to Ship In-Browser AI Apps With No Server and No Inference Bill
Microsoft Authenticator Android Flaw Shows Why Apps Must Verify Who Is Calling Them
Zero-Budget AI Apps Work Best as Self-Feeding Sites, Not Open Tools
A Useful Software Primitive Should Be Easy to Explain and Hard to Replace
Apple's US Link-Out Fee Is Zero Until a Judge Sets It, and Apple Is Asking for Up to 15%
20 Primitive Project Ideas: Tiny Go Engines That Do the Essential 10% of Redis, Kafka and Etcd
40 Projects Worth Building: SQLite for X, Tiny Infrastructure, MCP Tools and Edge Systems
40 Radically Different Infrastructure Primitives, From Reach and Expect to Ignorance and Stop

Media Partners

  • AltSql.com
  • Technology Conferences
  • API Coding
AltSql Is a Native Hybrid of Key-Value and SQL, From the Sensor to the Gateway
AltSql Core and AltSql DB Are Now Open Source Under the Apache License 2.0
Introduction
AltSql DB 0.3 Adds Secondary Indexes and Statement Savepoints: A Lookup on a Million Rows Ran 198 Times Faster Than a Full Scan
How It Works
AltSql DB 0.2: SQL and Direct Calls Wrote Byte-Identical Files Over 100,000 Random Steps
AltSql DB Keeps a Whole Fleet in One File and Reads by Key 5 Times Faster Than SQLite Through SQL
AltSql Core and Six Prototypes Built Around It
AltSql Ask Puts One SQL Question to 500 Simulated Machines and Brings Back Only the Answers
Machine Monitoring Case Study: Overheat Alarm Handled on the Sensor While Wi-Fi Was Down
electronica 2026, November 10-13, Messe München, Munich
KIOXIA Forum 2026, November 26, TODA Hall and Conference Tokyo, Tokyo
Black Hat Europe 2026, December 7-10, Excel London, London
ISC East 2026, November 3-5, Javits Center, New York
Cantor Convergence 2026, November 10-12, The Ritz-Carlton South Beach, Miami Beach
MATLAB EXPO Online 2026, November 4-5, Online
SGLang Summit 2026, November 12-13, Fort Mason Center for Arts and Culture, San Francisco
Visual Semiconductor and INCERGO at CES 2027, January 6-9, Las Vegas Convention Center, Las Vegas
Conferences Are Theatre and the Hallway Is the Product
SoftDevStudio.com at Web Summit 2026, November 9-12, MEO Arena and FIL, Lisbon
RFC 9457 Problem Details Gives API Errors One Shape Instead of Fifty Ad Hoc JSON Bodies
Idempotency Keys: How One Header Stops a Retry From Charging a Customer Twice
A Tiny ETL Binary Competes With curl, jq and SQLite in a Cron Job, So Build It That Small
An AI Agent Flight Recorder Belongs in One Portable File, the Way HAR Did It for HTTP
Benchmarking MCP Servers and Stateful API Workflows: Latency, Throughput and Tokens per Call
Diagnose a Slow API From One Request: DNS, Connect, TLS, Server Wait and Transfer
Finding API Endpoints, Tables and Flags Nobody Uses Takes Code and Traffic Together
Fuzzing MCP Servers: Generate Bad Arguments From the Tool Schema and Watch What Breaks
Give Any API a History: Poll It, Hash It and Query Old Versions With SQL
Infer Your API's Real Contract From Traffic, Then Diff It Against the Docs

Copyright © 2026 Technologies.org

Media Partners: Market Analysis · Market Research · Referently · Photography