Homerun, a quick commerce platform specializing in construction and finishing materials, has raised $12 million in a Series A+ funding round. Nexus Venture Partners led this round, with participation from existing investors: Sorin Investments, Titan Capital, Sparrow Capital, and Consumer Collective by Atrium.
The capital raised will be used by the company to expand its geographical presence, as well as for further investments in developing supply chains, technologies, and initiatives to expand product categories. Earlier this year, the startup had raised $6.6 million.
Homerun claims to provide contractors and homeowners with construction and finishing materials at competitive prices within 60 minutes. The platform focuses primarily on high-frequency categories, including cement, plywood, rebar, paints, and wires.
The startup reports completing 100,000 orders and demonstrating eightfold growth over the past year. Homerun's founder, Pukhraj Grewal, aims to solve long-standing issues related to unreliable supply, opaque pricing, sudden shortages, and labor downtime.
Pukhraj Grewal noted: 'Purchasing construction materials in India has remained too fragmented and inefficient for too long. At HomeRun, we are making it as reliable, transparent, and fast as any modern e-commerce experience. Strong customer response reinforces the scale of this opportunity.'
Previously, Pukhraj co-founded Tornado, a technology-driven interior construction startup, and Project Hero, a platform for hiring construction labor.
Anand Datta, a partner at Nexus Venture Partners, stated that 'our quick commerce concept focuses on formalizing supply chains while retaining the advantage of proximity of small local traders. Construction materials and the local apparatus are one such supply chain that has largely avoided digitalization. The company is growing rapidly without offering credit and possesses excellent unit economics.'
Convex has raised $57 million in a Series B funding round. These funds are planned to be used to scale the development of its backend platform, which is specifically designed for artificial intelligence. The financing was led by Insight Partners, with participation from Etna Labs and all previous investors, including a16z and Spark Capital.
Thanks to this Series B round, Convex's total raised capital has reached $110.5 million. The company stated that it intends to use this money to improve its core platform, enhance tools for agentic development, and increase its staff in its San Francisco office.
Founded in 2021 by former Dropbox infrastructure engineers, Convex offers developers a fully integrated backend platform. This platform combines databases, functions, workflows, search, synchronization, authentication, and file storage into a single solution built for software development.
The fundraising comes amid the trend towards AI-powered coding and autonomous development tools. Convex believes that developers need infrastructure capable of handling the increasing complexity of applications while maintaining a high level of reliability.
The company posits that traditional backend infrastructure struggles to support the growing use of AI coding agents. According to the company, AI-generated applications based on standard databases can subtly introduce data errors that only manifest after deployment.
Internal testing revealed that 90% of AI-built applications running on traditional databases encountered data corruption in real-world operating conditions. In comparison, applications developed on the Convex platform passed the same tests without any failures.
Convex's backend includes built-in ACID transactions, end-to-end type checking, real-time subscriptions, and TypeScript support. These features reduce the need for manual integration of multiple services for developers and increase application reliability. Furthermore, Convex packages search capabilities, workflow automation, authentication, file storage, and retrieval-augmented generation functions, allowing AI applications to be enriched with organization-specific knowledge.
The platform has gained significant traction among developers building AI-based software. Convex currently supports nearly two million applications created by about 500,000 developers worldwide. The company also reported over 1.2 million weekly downloads from npm.
Convex's clients include OpenAI, Tripadvisor, Solana, Zapier, and Reducto. These organizations use Convex to simplify backend development while ensuring scalable, reliable performance for their applications.
Insight Partners noted that the growing adoption of AI coding assistants has exposed weaknesses in traditional software infrastructure. The investment firm believes that Convex provides developers with a more robust foundation where backend components function cohesively without the need for complex configuration.
Convex plans to use the new funding to accelerate product innovations amidst the transformation of software engineering influenced by AI. The company will direct funds toward creating new tools that simplify application development for both developers and AI agents that build software.
The company will continue to provide highly reliable backend products supporting increasingly complex AI-based applications globally. Convex is convinced that stable backend infrastructure is becoming critically important as artificial intelligence generates large volumes of application code.
Instead of combining disparate components, Convex offers a single unified platform, which effectively reduces complexity and prevents software bugs. The raised funds will strengthen Convex's engineering team and significantly expand its global developer ecosystem and platform capabilities. With fresh capital and a growing customer base, Convex aims to solidify its position in the evolving AI development infrastructure market, making the creation of reliable software faster and easier for both human developers and AI agents.
The industry is showing increased interest in chip speed, but significant performance losses occur in the interaction between the chips themselves. In dense artificial intelligence clusters, GPUs often operate at only 30–40% capacity, meaning that most of the time powerful chips are idle waiting for data rather than processing it.
The startup Eliyan, which has received support from Cisco, raised $145 million with a company valuation of one billion dollars to address this issue. As cluster sizes increase, data transfer speed becomes insufficient. Simply increasing data volume through conventional copper wires leads to increased power consumption, heat generation, and ultimately, slower chip operation. Eliyan offers a solution to this problem.
Eliyan does not manufacture the chips itself; instead, it creates the connections between them. Founded in Santa Clara by Ramin Farjadrad, Patrick Soheil, and Cyrus Ziae, the company uses its NuLink PHY cores to move data between crystals on the same package and between separately packaged chips. Its NuGear chiplets are responsible for memory scaling and inter-rack scaling.
The most critical aspect for chip manufacturers is packaging. Many competing methods require the use of expensive silicon interposers to achieve high throughput. Eliyan's chiplets achieve comparable metrics using standard organic substrates, allowing chip designers to increase throughput without completely redesigning the chip's physical layout.
The funding round was led by Seligman Ventures, and Seligman Managing Partner Umesh Padwal joined Eliyan's board of directors. He was previously a member of the Mellanox board before its acquisition by Nvidia for $6.9 billion, an event that changed the industry's perception of networking technologies. Cisco and Lumentum are new strategic investors: Cisco has a division specializing in optics and silicon, and Lumentum manufactures lasers and optical modules that hyperscale data centers will purchase. In addition to these, Eliyan has received support from several leaders in the chip industry.
Prior to this round, Eliyan had investments from AMD, Arm, Meta, Coherent, Intel Capital, and Samsung Catalyst Fund. Investors who invest at such early stages typically aim to participate in shaping future technologies rather than just seeking profit. The company stated that it plans to begin shipping its first chiplets later this year and expects to reach hundreds of millions of dollars by 2027. Furthermore, the company holds over 100 patents.
Eliyan's main bet is that interconnectivity is the bottleneck for chiplets, and the company owns the intellectual property to solve it. This gives it an advantage over major players like Nvidia and Intel, who are beginning to focus on the same problem. The next phase of the AI race will be determined not by who releases the fastest chip, but by who can move data across the entire warehouse of such chips without exceeding the power budget, and it is in this race that Eliyan attracted $145 million.
