Pakistani-American Entrepreneur's AI Startup Commands $18 Billion Valuation
Anysphere, founded in 2022 by Sualeh Asif, Arvid Lunnemark, Aman Sanger and Michael Truell, has been approached by investors about a deal that would more than double its valuation in a new funding round to $18 to $20 billion, according to Bloomberg. Sualeh Asif, 25, the Chief Product Officer and co-founder, hails from Karachi, Pakistan. He studied at Nixor College in Karachi before going to the United States in 2018. He met the other three co-founders while studying at the Massachusetts Institute of Technology (MIT). One-half (Sualeh Asif and Aman Sanger) of the startup’s founding team is of South Asian origin. The company moved to San Francisco soon after its founding in 2022.
![]() |
Sualeh Asif, Anyshphere Inc's Pakistani-American Founder |
Anysphere’s lead product is an AI code editor called Cursor AI. It works as an AI assistant that writes code and also anticipates code developer's needs, identifies inefficiencies, and helps solve complex problems with relative ease. On his personal website sualehasif.me, Sualeh says: "I’m building Cursor to discover a new way to write code. I owe much of my fun to my friends and MIT. I am extremely excited about the new capabilities of LLMs and applications to code tools".
In early 2024, Anysphere raised $60 million in a Series A financing (co-led by Andreessen Horowitz and Thrive Capital and included Patrick Collison, CEO of Stripe) at a $400 million post-money valuation, according to TechCrunch. It was followed by a $105 million Series B funding round in December 2024. Earlier this month, Anysphere said it raised $900 million at a valuation of $9.9 billion, including the new investment, in a round led by Thrive Capital. Andreessen Horowitz, Accel and DST Global also participated.
Several investors have recently floated a funding offer that would value Anysphere at $18 billion to $20 billion, according to Bloomberg. The startup's revenue has recently surpassed $500 million on an annualized basis.
“We have not engaged with or spoken with any investors about financing since our Series C,” Chief Executive Officer Michael Truell wrote to Bloomberg in response to a query. “We’re not focused on fundraising and have lots to do on building the technology, product, and team. Cursor can be made much more useful, and we’re spending time on that.”
Related Links:
2021: A Banner Year For Tech Startups in Pakistan
Pakistan Projected to Be World's 6th Largest Economy By 2075
Digital Pakistan 2022: Broadband Penetration Soars to 90% of 15+ Population
Working Women Seeding a Silent Revolution in Pakistan
Socioeconomic Impact of New Infrastructure in Rural Pakistan
Pakistan Gets First Woman Supreme Court Judge
Pakistan at 75
Growing Presence of Pakistani Women in Science and Technology
Riaz Haq's Youtube Channel
Comments
I have developed in Kotlin, Java and Ruby using Cursor AI. It appears to support the most popular languages, I have not yet encountered a language it does not know. I asked it which languages it knows and it included Python and Prolog in the list. When prompted to solve a specific problem, Cursor will read through the code and reason about the solution by itself and make the code changes on your file system. It can then proceed to compile and run unit tests and programs. When it encounters build errors, it will try to fix them by itself. Through prompts, you can give it feedback to help it iterate on an acceptable solution. It can also write unit tests for your code by itself and take feedback. I asked it about its abilities to debug with Python and Prolog. It told me that it does not support interactive debugging specifically.
Python Debugging Support:
What Cursor AI can do:
Code analysis: Identify syntax errors, logical issues, and potential bugs through static analysis
Error explanation: Explain error messages and suggest fixes
Code review: Point out common programming mistakes and anti-patterns
Refactoring suggestions: Propose improvements to make code more robust
Test generation: Help create test cases to catch bugs
What it cannot do:
Runtime debugging: Cannot step through code execution, set breakpoints, or inspect variables at runtime
Interactive debugging: No debugger integration like pdb, ipdb, or IDE debuggers
Memory inspection: Cannot examine memory state or call stack during execution
I got a similar answer regarding its Prolog debugging abilities.
It is designed to be very similar to VSCode, a popular code editor for programmers. It uses all the same keyboard shortcuts. It is not necessarily like IntelliJ which does support runtime debugging for Java.
https://www.infoq.com/news/2025/06/void-ide-beta-release/
A new open-source AI-powered code editor, Void IDE, was recently released in beta, positioning itself as a privacy-focused and free alternative to popular closed-source AI editors like Cursor and GitHub Copilot. Backed by Y Combinator, Void IDE is a fork of Visual Studio Code. While Microsoft recently announced plans to open Source its GitHub Copilot Chat Extensionpossibly in a few months, the beta release is available now for the community to fiddle with.
The primary motivation behind Void IDE is to address concerns surrounding the privacy and cost associated with proprietary AI coding tools. Closed-source editors may require sending private code data through their backends, raising privacy issues and leading to ongoing subscription costs. While the backend may often keep only embeddings computed from the code, the original code may sometimes be recoverable. As the authors of the paper Mitigating Privacy Risks in LLM Embeddings from Embedding Inversion explain:
The surge in popularity of embedding vector databases in LLMs has been accompanied by significant concerns about privacy leakage. Embedding vector databases are particularly vulnerable to embedding inversion attacks, where adversaries can exploit the embeddings to reverse-engineer and extract sensitive information from the original text data.
Void IDE aims to provide options for keeping developers in control of their data. Void IDE can leverage various Large Language Models (LLMs), supporting direct integrations with services like Claude, GPT, and Gemini, as well as local model hosting via Ollama. This ensures that AI processing can happen locally or via direct API calls, avoiding a third-party middleman.
Void IDE offers a range of AI-centric features familiar to users of tools like Cursor. These include inline code editing, contextual AI chat, and code generation. The editor also features advanced capabilities like file system awareness for codebase-wide context and the ability to view/edit the underlying prompts sent to the AI.
Being a fork from VS Code, Void IDE is able to let its users migrate their themes, key bindings, and settings.
Developers on Hacker News and Reddit expressed interest, particularly regarding its open-source nature and privacy stance. Discussions include comparisons to other AI coding tools and editors. Some developers expressed skepticism about the proliferation of VS Code forks, with others asking why not use an extension instead. The project is actively developed, with the team encouraging contributions from the community to shape its future roadmap.