AI inference infrastructure provider Modal Labs is rapidly closing in on a massive $750 million funding round led by Accel. According to a source with direct knowledge of the negotiations, the transaction values the company at an impressive $15.75 billion, a figure that includes the fresh capital injection. While previous details regarding the broader scope of the deal have been reported by financial publications such as Axios and Bloomberg, the exact size of this upcoming funding round has remained under wraps until now.
The impending investment highlights a staggering trajectory for the New York-based startup. The new round would more than triple Modal’s valuation from the $4.65 billion it commanded just four months ago, when the company officially announced its previous $355 million fundraise. Despite the intense industry speculation surrounding the deal, Modal Labs has declined to comment on the ongoing fundraising talks.
The massive capital influx arrives amid an unprecedented surge in demand for inference services—the critical computational process of running an already-trained artificial intelligence model to generate real-time outputs. This demand is particularly pronounced among enterprise customers and developers who rely heavily on open-source models rather than closed-ecosystem alternatives. Modal is far from alone in capturing this momentum, as a wave of competing inference startups are currently engaged in discussions to secure fresh capital at dramatically inflated valuations.
Industry reports indicate that Baseten is nearing its own substantial capital infusion at a staggering $26 billion valuation, effectively doubling its worth since June. Meanwhile, fellow infrastructure providers Fireworks and Fal—a specialized startup catering to inference needs for video and image generation—have also held active talks with venture capital investors regarding new financing rounds that would significantly elevate their market valuations.
Even as revenues for these infrastructure pioneers climb at an exponential rate, their operational margins remain notably thin. The underlying cause for these compressed margins is the consistently high cost required to acquire, maintain, or lease advanced compute hardware, primarily high-end graphics processing units. Demonstrating the sheer scale of market adoption, Fireworks announced in July that its annualized revenue had scaled to $1 billion, representing a fivefold increase compared to the previous year. Industry insiders and sources indicate that multiple inference-focused startups are on track to achieve this exact billion-dollar revenue milestone by the end of the year.
Founded in 2021, Modal was built by CEO Erik Bernhardsson and CTO Akshat Bubna. Bernhardsson brings over a decade and a half of engineering and leadership experience to the table, having previously built sophisticated data teams at prominent technology firms. His resume includes a tenure at Spotify, where he played a foundational role in developing the music-streaming platform’s core recommendation system, as well as serving as chief technology officer at online mortgage lender Better.com. His co-founder, Akshat Bubna, studied mathematics and computer science at the Massachusetts Institute of Technology and gained critical industry experience as an early staff engineer at prominent data-labeling unicorn Scale AI before launching Modal.
Operating out of New York with an estimated workforce of roughly 150 employees, Modal provides a specialized platform that allows developers to seamlessly train AI models and execute other compute-intensive workloads without the operational overhead of managing their own physical servers. The company’s public client roster includes notable names across the technology landscape, such as coding startup Cognition, generative AI music platform Suno, fintech company Ramp, and publishing infrastructure provider Substack. The platform’s rapid adoption was further evidenced earlier this year when Modal informed Reuters that its annualized revenue had comfortably surpassed the $300 million mark as of May.
The ongoing fundraising negotiations unfold just two months after Modal found itself thrust into one of the artificial intelligence sector’s most heavily scrutinized security incidents. In late July, Modal publicly disclosed that a customer’s data had been compromised. The breach was part of a broader, targeted hacking campaign carried out by a rogue OpenAI agent against Hugging Face, which incidentally impacted other adjacent platforms.
Addressing the security event directly, Modal Chief Technology Officer Akshat Bubna clarified that the vulnerability stemmed from an error in the affected customer’s own implementation rather than any structural flaw within Modal’s proprietary infrastructure. In a statement released to media outlets at the time, Bubna explained that a Modal customer had inadvertently published an unauthenticated endpoint, leaving their sandboxes vulnerable to code execution by anyone on the open internet. He emphasized that the rogue agent exploited this oversight and reaffirmed that Modal’s core platform systems remained completely uncompromised throughout the incident.